CN101702197A - Method for detecting road traffic signs - Google Patents
Method for detecting road traffic signs Download PDFInfo
- Publication number
- CN101702197A CN101702197A CN200910154066A CN200910154066A CN101702197A CN 101702197 A CN101702197 A CN 101702197A CN 200910154066 A CN200910154066 A CN 200910154066A CN 200910154066 A CN200910154066 A CN 200910154066A CN 101702197 A CN101702197 A CN 101702197A
- Authority
- CN
- China
- Prior art keywords
- road signs
- geometric configuration
- road
- basic
- color
- 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
Images
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention discloses a method for detecting road traffic signs, which comprises the steps: providing color/shape pairs through the analysis of basic colors and geometrical shapes of the road traffic signs, constructs a relationship model between the colors and the geometrical shapes of the road traffic signs according to the color/shape pairs, and conducting the detection for the road traffic signs by utilizing the model, therefore, effectiveness and correctness of the detection for the road traffic signs are greatly enhanced; the model covering three types of 116 traffic signs in China is beneficial for the construction of a relatively perfect TSR system; the road traffic signs are preliminarily divided into seven subtypes of traffic signs while the detection, thus realizing rough classification of the road traffic signs. In this case, complexity of the road traffic signs recognition (TSR) system is effectively reduced and great advantages are achieved for the real-time property and the effectiveness of the TSR system; and furthermore, complexity of the information of the traffic signs can be simplified significantly, which is the foundation of accomplishing fast detection, and solving the problem of rough classification during the detection stage.
Description
Technical field
The present invention relates to a kind of recognition methods of road signs, especially relate to a kind of detection method of road signs.
Background technology
Intelligent transportation system (Intelligent Transportation System, ITS) be to propose for solving serious day by day urban traffic conditions, be that technology such as the detection of a collection, communication, control and computing machine are the integrated information system of one, become the sciemtifec and technical sphere of extensively being paid close attention at present.(Traffic Sign Recognition TSR) is the important component part of intelligent transportation system in road signs identification.Road signs identification mainly comprises two basic fundamental links: at first be the detection of traffic sign, comprise the location and the necessary pre-service of traffic sign; Next is the differentiation of traffic sign, comprises the feature extraction and the classification of traffic sign.Wherein, the detection of traffic sign is the key issue that will solve, and is the prerequisite that realizes the traffic sign correct decision.
Present traffic sign detects research mainly to be undertaken by two technology paths.At first, because colouring information is subjected to the influence of factor such as illumination easily, and the quantity of information that gray level image need be handled is less, therefore, the research that a lot of traffic signs detect is based on all that gray level image carries out, but based on the method for traffic sign detection of gray level image because the grey level of different colours differs very little sometimes, be difficult to distinguish, especially less or uneven illumination is even and when the interference of similar traffic sign feature is arranged when the contrast of outdoor scene traffic sign image, and this method all can increase misclassification rate.Secondly, the traffic sign that just is based on coloured image detects, and color is the important attribute of traffic sign, and with respect to the background area, its color mostly is distinct eye-catching, forms stronger color contrast with the peripheral region; Simultaneously, the information that coloured image provides is abundanter than gray level image, and therefore, along with the rapid raising of computer process ability, the research that the application color image processing carries out the road signs detection begins to increase.Method for traffic sign detection based on coloured image mainly contains following several:
(1) utilize the colouring information of traffic sign to carry out image segmentation, size, length breadth ratio and the position probing in image thereof according to traffic sign goes out traffic sign then;
(2) the CF information of application traffic sign detects traffic sign;
(3) utilize color space that the traffic sign scene graph is cut apart after, adopt genetic algorithm further to detect traffic sign.
Though above-mentioned several method for traffic sign detection based on coloured image has improved the detection effect of traffic sign, but detected object (being traffic sign) is generally more single, how with a certain class (as prohibitory sign or Warning Mark or warning notice) or only select several traffic signs in a certain class as detected object, when the sample number of detected object more for a long time, the detection accuracy of these methods obviously descends, thereby has reduced the discrimination of detected object.
Summary of the invention
Technical matters to be solved by this invention provides a kind of validity that can effectively improve the road signs detection, and detects the higher method for traffic sign detection of accuracy.
The present invention solves the problems of the technologies described above the technical scheme that is adopted: a kind of detection method of road signs, road signs have basic colors and basic two kinds of attributes of geometric configuration, described basic colors comprises redness, blueness, yellow, black and white, described basic geometric configuration comprises circle, rectangle, equilateral triangle, del and octagon, described road signs comprise prohibitory sign, Warning Mark and warning notice, and this detection method may further comprise the steps:
1. by analyzing the basic colors of road signs, determine with the redness to be that the road signs of basic colors are prohibitory sign, determine with the blueness to be that the road signs of basic colors are Warning Mark, determine with black to be that the road signs of basic colors are warning notice or are prohibitory sign; By analyzing the basic geometric configuration of road signs, determine with del to be that the road signs of basic geometric configuration are prohibitory sign, determine with octagon to be that the road signs of basic geometric configuration are prohibitory sign, determine with circle to be that the road signs of basic geometric configuration are prohibitory sign or are Warning Mark, determine with rectangle to be that the road signs of basic geometric configuration are Warning Mark, determine with equilateral triangle to be that the road signs of basic geometric configuration are warning notice;
2. the redness in the basic colors is designated as C
1, the blueness in the basic colors is designated as C
2, the black in the basic colors is designated as C
3, the del in the basic geometric configuration is designated as S
1, the octagon in the basic geometric configuration is designated as S
2, the circle in the basic geometric configuration is designated as S
3, the rectangle in the basic geometric configuration is designated as S
4, the equilateral triangle in the basic geometric configuration is designated as S
5, the set of note basic colors is VC, VC={C
1, C
2, C
3, the set of remembering basic geometric configuration is VS, VS={S
1, S
2, S
3, S
4, S
5; By i basic colors C
iWith j basic geometric configuration S
jConstitute the color shape to (C
i, S
j), wherein, i=1,2 ..., n, j=1,2 ..., m, n=3, m=5, all colours shape of note prohibitory sign is VCSP to the set that constitutes
1, VCSP
1={ (C
1, S
1), (C
1, C
2), (C
1, S
3), (C
3, S
3), all colours shape of note Warning Mark is VCSP to the set that constitutes
2, VCSP
2={ (C
2, S
3), (C
2, S
4), all colours shape of note warning notice is VCSP to the set that constitutes
3, VCSP
3={ (C
3, S
5), the right set of all colours shape of note road signs is VCSP
CN, VCSP
CN={ VCSP
1, VCSP
2, VCSP
3}={ (C
1, S
1), (C
1, S
2), (C
1, S
3), (C
3, S
3), (C
2, S
3), (C
2, S
4), (C
3, S
5); According to all colours shape of prohibitory sign to all colours shape of, Warning Mark to and all colours shape of warning notice right, the color of structure road signs and the relation of geometric configuration are expressed as
Wherein, TSR, TSG, TSW, TS are logical variable, there are road signs in the road traffic scene image that TS represents to obtain by camera head, TSR represents to have prohibitory sign in the road traffic scene image, TSG represents to have Warning Mark in the road traffic scene image, TSW represents to have prohibitory sign in the road traffic scene image, and TSG represents to have warning notice in the road traffic scene image; According to the color of road signs and the relation of geometric configuration, with the road signs rough sort is seven traffic sign subclasses, these seven traffic sign subclasses are respectively redness-del traffic sign, redness-octagon traffic sign, redness-circular traffic sign, black-circular traffic sign, blueness-circular traffic sign, blueness-rectangle traffic sign, black-equilateral triangle traffic sign, wherein, redness-del traffic sign, redness-octagon traffic sign, redness-circular traffic sign, black-circular traffic sign is a prohibitory sign, blueness-circular traffic sign, blueness-rectangle traffic sign is a Warning Mark, and black-equilateral triangle traffic sign is a warning notice;
3. obtain the road traffic scene image by camera head, defining the current road traffic scene image that obtains is current road traffic scene image, with the color space of current road traffic scene image from the RGB color notation conversion space to the HSI color space, 3 color components of RGB color space are respectively red R, green G and blue B, and 3 color components of HSI color space are respectively tone H, saturation degree S and intensity I;
4. according to the basic colors of road signs, adopt and current road traffic scene image is carried out color extracting, obtain bianry image based on the color extracting method of HSI color space;
5. adopt edge detection method that bianry image is carried out rim detection,, obtain doubtful road signs zone to remove noise spot;
6. all colours shape according to the basic geometric configuration of road signs and road signs is right, judges whether doubtful road signs zone is certain road signs zone; When doubtful road signs zone is defined as certain road signs zone, then according to the basic colors and the basic geometric configuration in the relation of the color of road signs and geometric configuration and certain road signs zone, determine the traffic sign subclass that certain road signs zone is affiliated, realize the detection of road signs; When doubtful road signs zone is defined as non-road signs zone, does not then handle and finish.
Described step in 3. with the detailed process of color space from the RGB color notation conversion space to the HSI color space of current road traffic scene image is: 3 color components to the RGB color space carry out normalized at first respectively, obtain 3 color components normalized value separately, the normalized value of note red R is r, the normalized value of remembering green G is g, the normalized value of remembering blue B is b
Utilize the normalized value b of the normalized value g of the normalized value r of red R, green G and blue B to calculate 3 color components of HSI color space respectively then,
Wherein,
Function arctan () is for asking tan, and function m in () is the function of minimizing.
4.-1, for the redness in the basic colors of road signs the detailed process based on the color extracting method of HSI color space that described step carries out to current road traffic scene image that color extracting adopted in 4. is:, the span of the value of the corresponding tone H of redness is set at [315 °, 360 °] ∪ [0 °, 25 °]; For the blueness in the basic colors of road signs, the span of the value of the corresponding tone H of blueness is set at [200 °, 260 °], and the span of the value of the corresponding saturation degree S of blueness is set at [0.2 ,+∞]; For the black in the basic colors of road signs, the span of the value of the corresponding saturation degree S of black is set at [0.2 ,+∞]; 4.-2, travel through each pixel of current road traffic scene image, whether the value of judging the tone H of pixel belongs to [315 °, 360 °] ∪ [0 °, 25 °], if, then the pixel value with this pixel is changed to 0, otherwise whether the value of the saturation degree S of judge whether the value of the tone H of pixel belongs to [200 °, 260 °] and pixel belongs to [0.2, + ∞], if then the pixel value with this pixel is changed to 0, otherwise, whether the value of judging the saturation degree S of pixel again belongs to [0.2, + ∞], if then the pixel value with this pixel is changed to 0, otherwise, the pixel value of this pixel is changed to 1; 4.-3, obtain bianry image.
The edge detection method of described step in 5. is expansion and the erosion operation in the mathematical morphology.
6.-1, utilize known Hough transformation to obtain the basic geometric configuration in doubtful road signs zone described step concrete steps 6. are:, judge then whether the basic geometric configuration in doubtful road signs zone belongs to the set VS of basic geometric configuration, if, then continue to carry out, otherwise, determine that this doubtful road signs zone is non-certain road signs, do not handle and finish; 6.-2, right according to the color geometric configuration of road signs, whether the binary that the basic colors of judging doubtful road signs zone and basic geometric configuration constitute to belonging to the right set VCSP of all colours shape of road signs
CN, if, determine that then this doubtful road signs zone be certain road signs zone, and continue to carry out, otherwise, determine these doubtful road signs regional be non-certain road signs zone, do not handle and finish; 6.-3, according to the basic colors and the basic geometric configuration in the relation of the color of road signs and geometric configuration and certain road signs zone, determine the traffic sign subclass that certain road signs zone is affiliated, realize the detection of road signs.
Compared with prior art, advantage of the present invention is as follows:
1. pass through to analyze the basic colors and the basic geometric configuration of road signs, propose the color shape to and according to the color shape to the color of having constructed road signs and the relational model of geometric configuration, utilize this model to carry out the detection of road signs, improved validity and detection accuracy that road signs detect greatly; This model has covered 116 kinds of traffic signs of all three major types of China, helps constructing a fairly perfect TSR system;
2. when detecting, road signs tentatively are divided into seven traffic sign subclasses, promptly realize the rough sort of road signs, like this, reduce the complicacy of road signs identifications (TSR) systems effectively, helped improving the real-time and the validity of TSR system;
3. the complicacy of abbreviation traffic sign information significantly, this is to realize fast detecting and the basis that solves the rough sort problem at detection-phase simultaneously;
4. because the inventive method has proposed the right notion of color shape, the basic colors of road signs and the unique deterministic dependence between the basic geometric configuration have promptly been considered, follow-up road signs are detected to detect road signs effectively, and can make the rough sort that in the process that detects, realizes road signs simultaneously, the color of the road signs of the inventive method structure and the relational model of geometric configuration provide good basis for effective detection of follow-up road signs.
Description of drawings
Fig. 1 is the part typical sample example of prohibitory sign, Warning Mark and warning notice;
Fig. 2 is the synoptic diagram that concerns of prohibitory sign, Warning Mark and warning notice and color;
Fig. 3 is the synoptic diagram that concerns of prohibitory sign, Warning Mark and warning notice and geometric configuration;
Fig. 4 is the basic framework synoptic diagram of unique definite relation between the color of road signs and the geometric configuration;
Fig. 5 is the base conditioning block diagram of the inventive method;
Fig. 6 a is the road traffic scene image of simulation;
Fig. 6 b is the bianry image that obtains after the color extracting;
Fig. 6 c is the doubtful road signs zone that obtains after the rim detection;
Fig. 6 d is for detecting the road signs that obtain.
Embodiment
Embodiment describes in further detail the present invention below in conjunction with accompanying drawing.
China directly traffic sign relevant with traffic safety has 116 (not comprising the traffic sign that can derive from), is divided into three major types, is respectively prohibitory sign, Warning Mark and warning notice.Wherein, 42 of prohibitory signs, 29 of Warning Marks, 45 of warning notices, Fig. 1 has provided the part typical sample example of three class signs.At present, detection to traffic sign has proposed two class methods, promptly based on the method for traffic sign detection of gray level image with based on the method for traffic sign detection of coloured image, wherein based on the method for traffic sign detection of coloured image in order to improve the validity that traffic sign detects, two prominent feature that made full use of traffic sign are color and geometry information, but another key character of having ignored traffic sign promptly has unique deterministic dependence between its color and the geometric configuration.
And the present invention is on the basis of the basic colors of having analyzed road signs and basic these two kinds of attributes of geometric configuration, the right notion of color shape has been proposed, the basic colors of road signs and the unique deterministic dependence between the basic geometric configuration have promptly been considered, follow-up road signs are detected to detect road signs effectively, and can make the rough sort that in the process that detects, realizes road signs simultaneously.On the right basis of color shape, the color of structure road signs and the relational model of geometric configuration, this model provides good guarantee for effective detection of follow-up road signs.Simultaneously on the basis of the relational model of the color of road signs and geometric configuration, a kind of new road signs detection method has been proposed, this method has covered 116 Chinese road signs of all three major types, when realizing detection, road signs tentatively are divided into seven traffic sign subclasses, promptly realize the rough sort of road signs, reduce the complicacy of road signs identifications (TSR) systems so effectively, improved the real-time and the validity of TSR system.
At first respectively the basic colors and the basic geometric configuration of all kinds of road signs are analyzed.From 116 road signs as can be known, these 116 road signs mainly are made of five kinds of basic colors, are respectively red, blue, yellow, black and white.Wherein, prohibitory sign is its basic colors with redness, and white is its background color, and the kernel pattern is based on black, and other has two special prohibitory signs is white background black; The basic colors of Warning Mark is blue, and the kernel pattern is based on white; Warning notice is its basic colors with yellow, and frame and kernel pattern thereof are black.The black content of considering 1/4th the sign kernel pattern of having an appointment in 45 warning notices is more, to such an extent as to yl moiety can not form the shape of a rule, shown in the warning notice W10 among Fig. 1.Therefore, if adopt the basic colors of yellow sign by way of caution to set up " relational model of color and geometric configuration ", will make the road signs detection method that the detection of warning notice is become very complicated.The outer rim of considering warning notice all is the black with certain width, and has the shape attribute of rule, so, the basic colors that the present invention adopts black to indicate by way of caution.
In addition, can know from Fig. 1 that most of road signs all have a very thin outer contour.But practice shows, takes in the road signs realistic picture that obtains, and these very thin outer contours are all fuzzyyer, even can't see, and therefore, these very thin outer contours can be ignored fully to the influence of road signs identification.1), be that the road signs of basic colors must be prohibitory sign with the redness so if ignore white, yellow and other secondary cause, then the available Fig. 2 of relation between color and the three major types road signs represents, specifically is expressed as follows:; 2) be that the road signs of basic colors must be Warning Mark with the blueness; 3), be the road signs of basic colors with black or be warning notice or be prohibitory sign.
116 road signs of China have five kinds of basic geometric configuratioies, are respectively circle, rectangle, equilateral triangle, del and octagon.Wherein, the quantity of circle, rectangle and equilateral triangle is more, is the main geometric configuration of three major types road signs, and del and octagon respectively have 1, and all are prohibitory signs, and all the other 40 prohibitory signs are circle; 29 Warning Marks have circular and two kinds of geometric configuratioies of rectangle; The basic geometric configuration of 45 warning notices is an equilateral triangle.1), circular or for prohibitory sign or be Warning Mark therefore, the relation between geometric configuration and the three major types road signs specifically is expressed as follows as shown in Figure 3:; 2), rectangle must be Warning Mark; 3), equilateral triangle must be warning notice; 4), del and octagon must be prohibitory sign.
Secondly, determine that the color shape is to reaching color shape pair set.Analysis according to above color attribute and geometric configuration attribute to 116 road signs of three major types, can draw an important conclusions: exist unique deterministic relation between the color of road signs and the geometric configuration really, the sign that is red circle, red del, red octagon and black circle is prohibitory sign, the sign of blue circle and blue rectangle is Warning Mark, the black equilateral triangle be masked as warning notice.Fig. 4 has provided the basic framework synoptic diagram of unique deterministic dependence between the color of road signs and the geometric configuration.
The color shape of the present invention's definition is as follows to reaching color shape pair set.
The definition that the color shape is right: note is taken the road signs real scene image zone that obtains and is A, and road signs real scene image zone A has n basic colors and m basic geometric configuration, remembers that i basic colors is C
i, remember that j basic geometric configuration is S
j, according to i the basic colors C of road signs real scene image zone A
iWith j basic geometric configuration S
j, make up a binary to (C
i, S
j), make CSP
k=(C
i, S
j) | k=1,2 ..., l, wherein, i=1,2 ..., n, j=1,2 ..., m, l=n * m determines that binary is to CSP
kFor the color shape of road signs real scene image zone A to (CSP, Color-Shape Pair).
The definition of color shape pair set: note is taken the road signs real scene image zone that obtains and is A, and road signs real scene image zone A has n basic colors and m basic geometric configuration, remembers that i basic colors is C
i, the set of note basic colors is VC, VC={C
1, C
2..., C
n, remember that j basic geometric configuration is S
j, the set of remembering basic geometric configuration is VS, VS={S
1, S
2..., S
m; I color attribute C according to road signs real scene image zone A
iWith j geometric configuration attribute S
j, the color shape that makes up among the A of road signs real scene image zone is right, is designated as CSP
k, CSP
k=(C
i, S
j) | k=1,2 ..., l, note is VCSP by l color shape to the color shape pair set of forming, VCSP={CSP
k| k=1,2 ..., l}, or VCSP={ (C
i, S
j) | i=1,2 ..., n, j=1,2 ..., m}, wherein, l=n * m.
If the color shape of utilizing the present invention to propose is right, then can when detecting, search out the effective coverage of satisfying traffic sign color and geometric match in follow-up traffic sign testing process.Below, it is right with the Warning Mark to be that example further specifies the color shape that the present invention proposes.At first, the basic colors of Warning Mark is blue, secondly, Warning Mark be shaped as circle or rectangle.Like this, Warning Mark should have fixing color shape to (blueness, circle) and (blueness, rectangle).The rest may be inferred, just can obtain with the color shape the color of expression and the relational model of geometric configuration.
At this, use C
1, C
2, C
3And C
4The basic colors of representing China's road signs respectively is red, blueness, black and other color, and other color comprises white and yellow; Use S
1, S
2, S
3, S
4, S
5And S
6The basic geometric configuration of representing China's road signs respectively is del, octagon, circle, rectangle, equilateral triangle and other shape, and other shapes comprise the different shape of road signs kernel pattern.So, according to top to road signs basic colors and basic geometric configuration analysis and the color shape to the definition of color shape pair set as can be known, the set of the basic colors of 116 road signs of China's three major types can be expressed as VC={C
1, C
2, C
3, the set of basic geometric configuration can be expressed as VS={S
1, S
2, S
3, S
4, S
5, the right set of all colours shape of note China road signs is VCSP
CN, the right set of all colours shape of note prohibitory sign is VCSP
1, the right set of all colours shape of note Warning Mark is VCSP
2, the right set of all colours shape of note warning notice is VCSP
3, wherein, VCSP
1, VCSP
2, VCSP
3Be VCSP
CNSubclass, can obtain VCSP according to analyzing
1={ (C
1, S
1), (C
1, S
2), (C
1, S
3), (C
3, S
3), VCSP
2={ (C
2, S
3), (C
2, S
4), VCSP
3={ (C
3, S
5), VCSP
CN={ VCSP
1, VCSP
2, VCSP
3}={ (C
1, S
1), (C
1, S
2), (C
1, S
3), (C
3, S
3), (C
2, S
3), (C
2, S
4), (C
3, S
5)
As the above analysis, China's road signs can be with 7 color shapes to describing, and wherein, prohibitory sign has 4, Warning Mark to have 2, warning notice to have 1 color shape right.If represent to have road signs in the road traffic scene graph with logical variable TS, represent to exist in the road traffic scene graph prohibitory sign, Warning Mark and warning notice respectively with TSR, TSG, TSW, then set up the color of road signs and the relation of geometric configuration (Color-Geometric Model, CGM), the relation of the color of road signs and geometric configuration can be expressed as with logical expression:
The color of these road signs and the relational model of geometric configuration show that 116 road signs of the three major types of China can be by 7 color shapes to representing that each color shape is to 1 subclass of expression road signs.Therefore, 116 road signs are divided into 7 subclasses, realized the rough sort of road signs, concrete classification situation is: (1), 42 prohibitory signs are divided into 4 subclasses: 38 red circle markers, 2 black circle markers, 1 red del sign, 1 octagon sign; These 4 subclasses correspond respectively to the color shape to (C
1, S
3), (C
3, S
3), (C
1, S
1), (C
1, S
2); (2), 29 Warning Marks are divided into 2 subclasses: 16 blue circle markers, 13 blue rectangle signs; These 2 subclasses correspond respectively to the color shape to (C
2, S
3), (C
2, S
4); (3), 45 warning notices have only 1 subclass, corresponding to the color shape to (C
3, S
5).
Because 116 road signs are divided into 7 subclasses, the sample number of each subclass greatly reduces, and therefore will improve the recognition efficiency and the recognition correct rate of road signs identification (TSR) system greatly.
On the basis of above-mentioned analysis, the present invention proposes a kind of detection method of road signs, as shown in Figure 5, this detection method may further comprise the steps:
1. by analyzing the basic colors of road signs, determine with the redness to be that the road signs of basic colors are prohibitory sign, determine with the blueness to be that the road signs of basic colors are Warning Mark, determine with black to be that the road signs of basic colors are warning notice or are prohibitory sign; By analyzing the basic geometric configuration of road signs, determine with del to be that the road signs of basic geometric configuration are prohibitory sign, determine with octagon to be that the road signs of basic geometric configuration are prohibitory sign, determine with circle to be that the road signs of basic geometric configuration are prohibitory sign or are Warning Mark, determine with rectangle to be that the road signs of basic geometric configuration are Warning Mark, determine with equilateral triangle to be that the road signs of basic geometric configuration are warning notice.
2. the redness in the basic colors is designated as C
1, the blueness in the basic colors is designated as C
2, the black in the basic colors is designated as C
3, the del in the basic geometric configuration is designated as S
1, the octagon in the basic geometric configuration is designated as S
2, the circle in the basic geometric configuration is designated as S
3, the rectangle in the basic geometric configuration is designated as S
4, the equilateral triangle in the basic geometric configuration is designated as S
5, the set of note basic colors is VC, VC={C
1, C
2, C
3, the set of remembering basic geometric configuration is VS, VS={S
1, S
2, S
3, S
4, S
5; By i basic colors C
iWith j basic geometric configuration S
jConstitute the color shape to (C
i, S
j), wherein, i=1,2 ..., n, j=1,2 ..., m, n=3, m=5, all colours shape of note prohibitory sign is VCSP to the set that constitutes
1, VCSP
1={ (C
1, S
1), (C
1, S
2), (C
1, S
3), (C
3, S
3), all colours shape of note Warning Mark is VCSP to the set that constitutes
2, VCSP
2={ (C
2, S
3), (C
2, S
4), all colours shape of note warning notice is VCSP to the set that constitutes
3, VCSP
3={ (C
3, S
5), the right set of all colours shape of note road signs is VCSP
CN, VCSP
CN={ VCSP
1, VCSP
2, VCSP
3}={ (C
1, S
1), (C
1, S
2), (C
1, S
3), (C
3, S
3), (C
2, S
3), (C
2, S
4), (C
3, S
5); According to all colours shape of prohibitory sign to all colours shape of, Warning Mark to and all colours shape of warning notice right, the color of structure road signs and the relation of geometric configuration are expressed as
Wherein, TSR, TSG, TSW, TS are logical variable, there are road signs in the road traffic scene image that TS represents to obtain by camera head, TSR represents to have prohibitory sign in the road traffic scene image, TSG represents to have Warning Mark in the road traffic scene image, TSW represents to have prohibitory sign in the road traffic scene image, and TSG represents to have warning notice in the road traffic scene image; According to the color of road signs and the relation of geometric configuration, with the road signs rough sort is seven traffic sign subclasses, these seven traffic sign subclasses are respectively redness-del traffic sign, redness-octagon traffic sign, redness-circular traffic sign, black-circular traffic sign, blueness-circular traffic sign, blueness-rectangle traffic sign, black-equilateral triangle traffic sign, wherein, redness-del traffic sign, redness-octagon traffic sign, redness-circular traffic sign, black-circular traffic sign is a prohibitory sign, blueness-circular traffic sign, blueness-rectangle traffic sign is a Warning Mark, and black-equilateral triangle traffic sign is a warning notice.
3. obtain the road traffic scene image by camera head, defining the current road traffic scene image that obtains is current road traffic scene image, with the color space of current road traffic scene image from the RGB color notation conversion space to the HSI color space, 3 color components of RGB color space are respectively red R, green G and blue B, and 3 color components of HSI color space are respectively tone H, saturation degree S and intensity I.Because the HSI color space can embody the human vision feature, and the color attribute and the geometric configuration attribute of the road traffic scene image represented of HSI color space are more obvious, are easy to follow-up color extracting, rim detection etc.
At this, with the detailed process of color space from the RGB color notation conversion space to the HSI color space of current road traffic scene image be: 3 color components to the RGB color space carry out normalized at first respectively, obtain 3 color components normalized value separately, the normalized value of note red R is r, the normalized value of remembering green G is g, the normalized value of remembering blue B is b
Utilize the normalized value b of the normalized value g of the normalized value r of red R, green G and blue B to calculate 3 color components of HSI color space respectively then,
Wherein,
Function arctan () is for asking tan, and function m in () is the function of minimizing.
4. according to the basic colors of road signs, adopt and current road traffic scene image is carried out color extracting, can obtain bianry image based on the color extracting method of HSI color space.
At this, based on the color extracting method of HSI color space is to determine the span of the color component of HSI color space according to the basic colors of road signs, each pixel in the current road traffic scene image is represented by three color components of HIS color space, the basic colors of road signs has redness, blueness and black, only used two color components of tone H and saturation degree S in concrete color extracting process, the tone H of the basic colors correspondence of road signs and the span of saturation degree S are as shown in table 1.4.-1, for the redness in the basic colors of road signs at this, the detailed process of color extracting method is:, the span of the value of the corresponding tone H of redness is set at [315 °, 360 °] ∪ [0 °, 25 °]; For the blueness in the basic colors of road signs, the span of the value of the corresponding tone H of blueness is set at [200 °, 260 °], and the span of the value of the corresponding saturation degree S of blueness is set at [0.2 ,+∞]; For the black in the basic colors of road signs, the span of the value of the corresponding saturation degree S of black is set at [0.2 ,+∞]; 4.-2, travel through each pixel of current road traffic scene image, whether the value of judging the tone H of pixel belongs to [315 °, 360 °] ∪ [0 °, 25 °], if, then the pixel value with this pixel is changed to 0, otherwise whether the value of the saturation degree S of judge whether the value of the tone H of pixel belongs to [200 °, 260 °] and pixel belongs to [0.2, + ∞], if then the pixel value with this pixel is changed to 0, otherwise, whether the value of judging the saturation degree S of pixel again belongs to [0.2, + ∞], if then the pixel value with this pixel is changed to 0, otherwise, the pixel value of this pixel is changed to 1; 4.-3, obtain bianry image.
The tone H of the basic colors correspondence of table 1 road signs and the span table of saturation degree S
| Basic colors | The H value | The S value |
| Red | ??[315°，360°]∪[0°，25°] | ??- |
| Blue | ??[200°，260°] | ??(0.2，+∞) |
| Black | ??- | ??(0.2，+∞) |
"-" expression is not considered in the table 1.
5. adopt edge detection method that bianry image is carried out rim detection,, obtain doubtful road signs zone to remove noise spot.
Rim detection claims edge extracting again, and relevant concrete grammar has a lot, is example to adopt expansion and erosion operation in the mathematical morphology in this specific embodiment, and corrosion is the dual operations of expansion, but they are not inverse operations, and corresponding formulas is as follows:
A. two-value expands
If two operands of vectorial addition come from X and B respectively, and get possible combination, then dilation operation arbitrarily
May be defined as:
B. two-value corrosion
If two operands of vector subtraction come from X and B respectively, and get possible combination, then erosion operation arbitrarily
May be defined as:
At this, utilize expansion in the morphology and erosion operation that bianry image is corroded and expand, the number of times that carries out with corrosion of expanding depends on the quantity of effective pixel points, is a variable.
6. all colours shape according to the basic geometric configuration of road signs and road signs is right, judge whether doubtful road signs zone is certain road signs zone, when doubtful road signs zone is defined as certain road signs zone, then according to the basic colors and the basic geometric configuration in the relation of the color of road signs and geometric configuration and certain road signs zone, determine the traffic sign subclass that certain road signs zone is affiliated, realize the detection of road signs; When doubtful road signs zone is defined as non-road signs zone, does not then handle and finish.Concrete steps are:
6.-1, utilize known Hough transformation to obtain the basic geometric configuration in doubtful road signs zone, judge then whether the basic geometric configuration in doubtful road signs zone belongs to the set VS of basic geometric configuration, if, then continue to carry out, otherwise, determine that this doubtful road signs zone is non-certain road signs, do not handle and finish; 6.-2, right according to the color geometric configuration of road signs, whether the binary that the basic colors of judging doubtful road signs zone and basic geometric configuration constitute to belonging to the right set VCSP of all colours shape of road signs
CN, if, determine that then this doubtful road signs zone be certain road signs zone, and continue to carry out, otherwise, determine these doubtful road signs regional be non-certain road signs zone, do not handle and finish; 6.-3, according to the basic colors and the basic geometric configuration in the relation of the color of road signs and geometric configuration and certain road signs zone, determine the traffic sign subclass that certain road signs zone is affiliated, realize the detection and the rough sort of road signs.
Above-mentioned steps 6.-1 in for the geometric configuration method of discrimination that can adopt existing any maturation that obtains of the basic geometric configuration in doubtful road signs zone, wherein, effect with Hough transformation is better, therefore adopts Hough transformation to realize the differentiation of geometric configuration in this specific embodiment.
After utilizing the inventive method to carry out the road signs detection, the result who detects is actually the result of a rough sort, subsequent treatment for each the traffic sign subclass after the rough sort all can be handled by a sorter or arbiter, and the complexity difference.For example, traffic sign subclass 1 and traffic sign subclass 2 are owing to all only comprised road signs, so subsequent treatment is very simple, as long as an arbiter, purpose is the result that the filtering flase drop is surveyed.And traffic sign subclass 7 is just different, has 45 road signs, has contained whole one big class road signs---warning notice, and therefore, the subsequent treatment more complicated of traffic sign subclass 7 need further be carried out disaggregated classification and differentiation.
Below for the feasibility and the realizability of the inventive method are described by emulation experiment.
Because it is relatively more difficult to obtain the actual scene image of whole 116 road signs,, mainly be to have adopted the road traffic scene image of a large amount of simulations therefore as experimental subjects except adopting segment path traffic scene image as the experimental subjects.The road traffic scene image of so-called simulation, the test pattern (figure that comprises torsional deformation) with 116 road signs sticks on the road signs that replace real scene shooting on the different real scene shooting road traffic scene images respectively exactly.The data of the road traffic scene image of simulation have 4 groups, wherein 1 group is directly to adopt the road signs of standard to make, other 3 groups is simulation geometric distortion, the road traffic scene image of the simulation that the standard road traffic sign that promptly adopts torsion resistance to be respectively 5 °, 10 ° and 15 ° is made.Every group of data all have 116 road signs, have the road traffic scene image of 464 width of cloth simulation.
Fig. 6 a has provided the road traffic scene image of width of cloth simulation, employing is carried out color extracting based on the color extracting method of HSI color space to the image shown in Fig. 6 a, the bianry image that obtains is shown in Fig. 6 b, but often there is a lot of noise spots owing to extract the bianry image that obtains, from Fig. 6 b also as can be seen, for removing the noise spot in the image shown in Fig. 6 b, need the image shown in Fig. 6 b is carried out rim detection, because the present invention has mainly adopted corrosion and dilation operation to realize rim detection, at this, after expanding, corrosion obtains doubtful road signs zone shown in Fig. 6 c.Image shown in Fig. 6 c is being carried out basic geometric configuration when judging, owing to consider the corner features of the basic geometric configuration of the road signs among the set VS of basic geometric configuration, therefore in emulation experiment, utilize existing Corner boarding matching method to determine the basic geometric configuration in doubtful road signs zone, color geometric configuration according to road signs is right then, and whether the binary that the basic colors of judging doubtful road signs zone and basic geometric configuration constitute to belonging to the right set VCSP of all colours shape of road signs
CNImage shown in Fig. 6 b satisfies condition after judging, again according to the basic colors and the basic geometric configuration in the relational model of the color of road signs and geometric configuration and certain road signs zone, determine the traffic sign subclass that certain road signs zone is affiliated, realize the detection of road signs, the result that detection obtains is shown in Fig. 6 d.
Process and result from above-mentioned emulation experiment, the accuracy that adopts the inventive method to carry out the detection of road signs can reach 100%, proof the inventive method has good accuracy, robustness and real-time, thereby the feasibility and the realizability of the inventive method have been proved, testing result shows that the inventive method has not only realized the rough sort of three major types road signs simultaneously simultaneously, and directly 116 road signs of this three major types is divided into seven traffic sign subclasses.
In order further to verify the validity of the inventive method, 4 groups of road signs have been carried out disaggregated classification (promptly differentiate) research, 4 groups of road signs promptly wherein 1 group be that the road signs of standard and other 3 groups are to adopt torsion resistance to be respectively the road signs of 5 °, 10 ° and 15 °.The object of handling without the inventive method is 116 road signs, the corresponding arbiter of each road signs, and the object after the inventive method is handled is seven traffic sign subclasses, the corresponding arbiter of each traffic sign subclass, the inventive method is more effective to the differentiation of road signs as can be known, and employed arbiter also still less.At this, adopt the support vector machine network as judgement system, respectively existing method for traffic sign detection based on coloured image is detected result who obtains and the result that the inventive method detection obtains and differentiate, the accuracy of differentiation is as shown in table 2.
Table 2 adopts the support vector machine network respectively the accuracy that the existing result who obtains based on method for traffic sign detection and the inventive method detection of coloured image differentiates to be compared as judgement system
The inventive method can provide for the differentiation of follow-up road signs well and ensure as can be seen from Table 2, has improved the accuracy of differentiating greatly.
Claims (5)
1. the detection method of road signs, road signs have basic colors and basic two kinds of attributes of geometric configuration, described basic colors comprises redness, blueness, yellow, black and white, described basic geometric configuration comprises circle, rectangle, equilateral triangle, del and octagon, described road signs comprise prohibitory sign, Warning Mark and warning notice, it is characterized in that this detection method may further comprise the steps:
1. by analyzing the basic colors of road signs, determine with the redness to be that the road signs of basic colors are prohibitory sign, determine with the blueness to be that the road signs of basic colors are Warning Mark, determine with black to be that the road signs of basic colors are warning notice or are prohibitory sign; By analyzing the basic geometric configuration of road signs, determine with del to be that the road signs of basic geometric configuration are prohibitory sign, determine with octagon to be that the road signs of basic geometric configuration are prohibitory sign, determine with circle to be that the road signs of basic geometric configuration are prohibitory sign or are Warning Mark, determine with rectangle to be that the road signs of basic geometric configuration are Warning Mark, determine with equilateral triangle to be that the road signs of basic geometric configuration are warning notice;
2. the redness in the basic colors is designated as C
1, the blueness in the basic colors is designated as C
2, the black in the basic colors is designated as C
3, the del in the basic geometric configuration is designated as S
1, the octagon in the basic geometric configuration is designated as S
2, the circle in the basic geometric configuration is designated as S
3, the rectangle in the basic geometric configuration is designated as S
4, the equilateral triangle in the basic geometric configuration is designated as S
5, the set of note basic colors is VC, VC={C
1, C
2, C
3, the set of remembering basic geometric configuration is VS, VS={S
1, S
2, S
3, S
4, S
5; By i basic colors C
iWith j basic geometric configuration S
jConstitute the color shape to (C
i, S
j), wherein, i=1,2 ..., n, j=1,2 ..., m, n=3, m=5, all colours shape of note prohibitory sign is VCSP to the set that constitutes
1, VCSP
1={ (C
1, S
1), (C
1, S
2), (C
1, S
3), (C
3, S
3), all colours shape of note Warning Mark is VCSP to the set that constitutes
2, VCSP
2={ (C
2, S
3), (C
2, S
4), all colours shape of note warning notice is VCSP to the set that constitutes
3, VCSP
3={ (C
3, S
5), the right set of all colours shape of note road signs is VCSP
CN, VCSP
CN={ VCSP
1, VCSP
2, VCSP
3}={ (C
1, S
1), (C
1, S
2), (C
1, S
3), (C
3, S
3), (C
2, S
3), (C
2, S
4), (C
3, S
5); According to all colours shape of prohibitory sign to all colours shape of, Warning Mark to and all colours shape of warning notice right, the color of structure road signs and the relation of geometric configuration are expressed as
Wherein, TSR, TSG, TSW, TS are logical variable, there are road signs in the road traffic scene image that TS represents to obtain by camera head, TSR represents to have prohibitory sign in the road traffic scene image, TSG represents to have Warning Mark in the road traffic scene image, TSW represents to have prohibitory sign in the road traffic scene image, and TSG represents to have warning notice in the road traffic scene image; According to the color of road signs and the relation of geometric configuration, with the road signs rough sort is seven traffic sign subclasses, these seven traffic sign subclasses are respectively redness-del traffic sign, redness-octagon traffic sign, redness-circular traffic sign, black-circular traffic sign, blueness-circular traffic sign, blueness-rectangle traffic sign, black-equilateral triangle traffic sign, wherein, redness-del traffic sign, redness-octagon traffic sign, redness-circular traffic sign, black-circular traffic sign is a prohibitory sign, blueness-circular traffic sign, blueness-rectangle traffic sign is a Warning Mark, and black-equilateral triangle traffic sign is a warning notice;
3. obtain the road traffic scene image by camera head, defining the current road traffic scene image that obtains is current road traffic scene image, with the color space of current road traffic scene image from the RGB color notation conversion space to the HSI color space, 3 color components of RGB color space are respectively red R, green G and blue B, and 3 color components of HSI color space are respectively tone H, saturation degree S and intensity I;
4. according to the basic colors of road signs, adopt and current road traffic scene image is carried out color extracting, obtain bianry image based on the color extracting method of HSI color space;
5. adopt edge detection method that bianry image is carried out rim detection,, obtain doubtful road signs zone to remove noise spot;
6. all colours shape according to the basic geometric configuration of road signs and road signs is right, judge whether doubtful road signs zone is certain road signs zone, when doubtful road signs zone is defined as certain road signs zone, then according to the basic colors and the basic geometric configuration in the relation of the color of road signs and geometric configuration and certain road signs zone, determine the traffic sign subclass that certain road signs zone is affiliated, realize the detection of road signs; When doubtful road signs zone is defined as non-road signs zone, does not then handle and finish.
2. the detection method of a kind of road signs according to claim 1, it is characterized in that during described step 3. the detailed process of color space from the RGB color notation conversion space to the HSI color space of current road traffic scene image being: 3 color components to the RGB color space carry out normalized at first respectively, obtain 3 color components normalized value separately, the normalized value of note red R is r, the normalized value of remembering green G is g, the normalized value of remembering blue B is b
Utilize the normalized value b of the normalized value g of the normalized value r of red R, green G and blue B to calculate 3 color components of HSI color space respectively then,
Wherein,
Function arctan () is for asking tan, and function m in () is the function of minimizing.
3. the detection method of a kind of road signs according to claim 1 and 2,4.-1, it is characterized in that during described step 4. the detailed process that current road traffic scene image carries out that color extracting adopted being: for the redness in the basic colors of road signs based on the color extracting method of HSI color space, the span of the value of the corresponding tone H of redness is set at [315 °, 360 °] ∪ [0 °, 25 °]; For the blueness in the basic colors of road signs, the span of the value of the corresponding tone H of blueness is set at [200 °, 260 °], and the span of the value of the corresponding saturation degree S of blueness is set at [0.2 ,+∞]; For the black in the basic colors of road signs, the span of the value of the corresponding saturation degree S of black is set at [0.2 ,+∞]; 4.-2, travel through each pixel of current road traffic scene image, whether the value of judging the tone H of pixel belongs to [315 °, 360 °] ∪ [0 °, 25 °], if, then the pixel value with this pixel is changed to 0, otherwise whether the value of the saturation degree S of judge whether the value of the tone H of pixel belongs to [200 °, 260 °] and pixel belongs to [0.2, + ∞], if then the pixel value with this pixel is changed to 0, otherwise, whether the value of judging the saturation degree S of pixel again belongs to [0.2, + ∞], if then the pixel value with this pixel is changed to 0, otherwise, the pixel value of this pixel is changed to 1; 4.-3, obtain bianry image.
4. the detection method of a kind of road signs according to claim 3 is characterized in that edge detection method during described step 5. is expansion and the erosion operation in the mathematical morphology.
5. the detection method of a kind of road signs according to claim 4,6.-1, utilize known Hough transformation to obtain the basic geometric configuration in doubtful road signs zone it is characterized in that described step concrete steps 6. are:, judge then whether the basic geometric configuration in doubtful road signs zone belongs to the set VS of basic geometric configuration, if, then continue to carry out, otherwise, determine that this doubtful road signs zone is non-certain road signs, do not handle and finish; 6.-2, right according to the color geometric configuration of road signs, whether the binary that the basic colors of judging doubtful road signs zone and basic geometric configuration constitute to belonging to the right set VCSP of all colours shape of road signs
CN, if, determine that then this doubtful road signs zone be certain road signs zone, and continue to carry out, otherwise, determine these doubtful road signs regional be non-certain road signs zone, do not handle and finish; 6.-3, according to the basic colors and the basic geometric configuration in the relation of the color of road signs and geometric configuration and certain road signs zone, determine the traffic sign subclass that certain road signs zone is affiliated, realize the detection of road signs.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910154066A CN101702197A (en) | 2009-10-26 | 2009-10-26 | Method for detecting road traffic signs |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910154066A CN101702197A (en) | 2009-10-26 | 2009-10-26 | Method for detecting road traffic signs |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101702197A true CN101702197A (en) | 2010-05-05 |
Family
ID=42157112
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200910154066A Pending CN101702197A (en) | 2009-10-26 | 2009-10-26 | Method for detecting road traffic signs |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101702197A (en) |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011140679A1 (en) * | 2010-05-13 | 2011-11-17 | 中国科学院自动化研究所 | Method for classifying shape images based on the topological theory of perceptual organization |
| CN102376166A (en) * | 2010-08-23 | 2012-03-14 | 福特全球技术公司 | Arrangement and method for recognizing road signs in motor vehicle |
| CN102819728A (en) * | 2012-07-17 | 2012-12-12 | 中国航天科工集团第三研究院第八三五七研究所 | Traffic sign detection method based on classification template matching |
| EP2573711A2 (en) | 2011-09-23 | 2013-03-27 | Ricoh Company, Ltd. | Traffic sign detecting method and traffic sign detecing device |
| EP2575077A2 (en) | 2011-09-29 | 2013-04-03 | Ricoh Company, Ltd. | Road sign detecting method and road sign detecting apparatus |
| CN103142187A (en) * | 2013-03-18 | 2013-06-12 | 慈溪迈思特电子科技有限公司 | Work place recognition system of automatic dust collector |
| CN103186905A (en) * | 2011-12-28 | 2013-07-03 | 现代自动车株式会社 | Color detector for vehicle |
| CN103577809A (en) * | 2013-11-12 | 2014-02-12 | 北京联合大学 | Ground traffic sign real-time detection method based on intelligent driving |
| CN103745481A (en) * | 2014-01-27 | 2014-04-23 | 哈尔滨工业大学 | Circle Hough transform high-speed VLSI (very-large-scale integration) structure |
| CN103986882A (en) * | 2014-05-21 | 2014-08-13 | 福建歌航电子信息科技有限公司 | Method for image classification, transmission and processing in real-time monitoring system |
| CN104091326A (en) * | 2014-06-16 | 2014-10-08 | 小米科技有限责任公司 | Method and device for icon segmentation |
| CN104299221A (en) * | 2014-07-11 | 2015-01-21 | 奇瑞汽车股份有限公司 | Determination method and apparatus for traffic sign image |
| CN104766042A (en) * | 2014-01-06 | 2015-07-08 | 现代摩比斯株式会社 | Method and apparatus for and recognizing traffic sign board |
| CN104766046A (en) * | 2015-02-06 | 2015-07-08 | 哈尔滨工业大学深圳研究生院 | Detection and recognition algorithm conducted by means of traffic sign color and shape features |
| CN105243664A (en) * | 2015-10-08 | 2016-01-13 | 东南大学 | Vision-based wheeled mobile robot fast target tracking method |
| CN105740877A (en) * | 2014-12-09 | 2016-07-06 | 比亚迪股份有限公司 | Traffic sign recognition method and device, and vehicle |
| CN106503607A (en) * | 2015-09-07 | 2017-03-15 | 北京易车互联信息技术有限公司 | Traffic sign recognition method and device |
| CN107103606A (en) * | 2017-02-27 | 2017-08-29 | 口碑控股有限公司 | A kind of image-recognizing method and device |
| CN109214434A (en) * | 2018-08-20 | 2019-01-15 | 上海萃舟智能科技有限公司 | A kind of method for traffic sign detection and device |
| CN110851349A (en) * | 2019-10-10 | 2020-02-28 | 重庆金融资产交易所有限责任公司 | Page abnormal display detection method, terminal equipment and storage medium |
| CN112789654A (en) * | 2018-10-04 | 2021-05-11 | 日立安斯泰莫株式会社 | Camera device |
-
2009
- 2009-10-26 CN CN200910154066A patent/CN101702197A/en active Pending
Cited By (33)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8732172B2 (en) | 2010-05-13 | 2014-05-20 | Institute Of Automation, Chinese Academy Of Sciences | Shape classification method based on the topological perceptual organization theory |
| WO2011140679A1 (en) * | 2010-05-13 | 2011-11-17 | 中国科学院自动化研究所 | Method for classifying shape images based on the topological theory of perceptual organization |
| CN102376166A (en) * | 2010-08-23 | 2012-03-14 | 福特全球技术公司 | Arrangement and method for recognizing road signs in motor vehicle |
| CN102376166B (en) * | 2010-08-23 | 2016-08-03 | 福特全球技术公司 | For identifying the apparatus and method of road signs in motor vehicles |
| CN103020623A (en) * | 2011-09-23 | 2013-04-03 | 株式会社理光 | Traffic sign detection method and equipment |
| EP2573711A2 (en) | 2011-09-23 | 2013-03-27 | Ricoh Company, Ltd. | Traffic sign detecting method and traffic sign detecing device |
| CN103020623B (en) * | 2011-09-23 | 2016-04-06 | 株式会社理光 | Method for traffic sign detection and road traffic sign detection equipment |
| US8983136B2 (en) | 2011-09-23 | 2015-03-17 | Ricoh Company, Ltd. | Traffic sign detecting method and traffic sign detecting device |
| EP2575077A2 (en) | 2011-09-29 | 2013-04-03 | Ricoh Company, Ltd. | Road sign detecting method and road sign detecting apparatus |
| CN103186905A (en) * | 2011-12-28 | 2013-07-03 | 现代自动车株式会社 | Color detector for vehicle |
| CN102819728A (en) * | 2012-07-17 | 2012-12-12 | 中国航天科工集团第三研究院第八三五七研究所 | Traffic sign detection method based on classification template matching |
| CN103142187A (en) * | 2013-03-18 | 2013-06-12 | 慈溪迈思特电子科技有限公司 | Work place recognition system of automatic dust collector |
| CN103577809A (en) * | 2013-11-12 | 2014-02-12 | 北京联合大学 | Ground traffic sign real-time detection method based on intelligent driving |
| CN103577809B (en) * | 2013-11-12 | 2016-08-17 | 北京联合大学 | A kind of method that traffic above-ground mark based on intelligent driving detects in real time |
| CN104766042B (en) * | 2014-01-06 | 2018-12-07 | 现代摩比斯株式会社 | Traffic sign boards recognition methods and device |
| CN104766042A (en) * | 2014-01-06 | 2015-07-08 | 现代摩比斯株式会社 | Method and apparatus for and recognizing traffic sign board |
| CN103745481A (en) * | 2014-01-27 | 2014-04-23 | 哈尔滨工业大学 | Circle Hough transform high-speed VLSI (very-large-scale integration) structure |
| CN103986882A (en) * | 2014-05-21 | 2014-08-13 | 福建歌航电子信息科技有限公司 | Method for image classification, transmission and processing in real-time monitoring system |
| CN104091326A (en) * | 2014-06-16 | 2014-10-08 | 小米科技有限责任公司 | Method and device for icon segmentation |
| CN104299221A (en) * | 2014-07-11 | 2015-01-21 | 奇瑞汽车股份有限公司 | Determination method and apparatus for traffic sign image |
| CN104299221B (en) * | 2014-07-11 | 2017-07-11 | 奇瑞汽车股份有限公司 | A kind of determination method and device of Traffic Sign Images |
| CN105740877A (en) * | 2014-12-09 | 2016-07-06 | 比亚迪股份有限公司 | Traffic sign recognition method and device, and vehicle |
| CN104766046A (en) * | 2015-02-06 | 2015-07-08 | 哈尔滨工业大学深圳研究生院 | Detection and recognition algorithm conducted by means of traffic sign color and shape features |
| CN104766046B (en) * | 2015-02-06 | 2018-02-16 | 哈尔滨工业大学深圳研究生院 | One kind is detected using traffic mark color and shape facility and recognition methods |
| CN106503607A (en) * | 2015-09-07 | 2017-03-15 | 北京易车互联信息技术有限公司 | Traffic sign recognition method and device |
| CN106503607B (en) * | 2015-09-07 | 2020-02-07 | 北京易车互联信息技术有限公司 | Road traffic sign identification method and device |
| CN105243664B (en) * | 2015-10-08 | 2018-07-17 | 东南大学 | A kind of wheeled mobile robot fast-moving target tracking method of view-based access control model |
| CN105243664A (en) * | 2015-10-08 | 2016-01-13 | 东南大学 | Vision-based wheeled mobile robot fast target tracking method |
| CN107103606A (en) * | 2017-02-27 | 2017-08-29 | 口碑控股有限公司 | A kind of image-recognizing method and device |
| CN109214434A (en) * | 2018-08-20 | 2019-01-15 | 上海萃舟智能科技有限公司 | A kind of method for traffic sign detection and device |
| CN112789654A (en) * | 2018-10-04 | 2021-05-11 | 日立安斯泰莫株式会社 | Camera device |
| CN110851349A (en) * | 2019-10-10 | 2020-02-28 | 重庆金融资产交易所有限责任公司 | Page abnormal display detection method, terminal equipment and storage medium |
| CN110851349B (en) * | 2019-10-10 | 2023-12-26 | 岳阳礼一科技股份有限公司 | Page abnormity display detection method, terminal equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101702197A (en) | Method for detecting road traffic signs | |
| CN108108761B (en) | Rapid traffic signal lamp detection method based on deep feature learning | |
| CN102968646B (en) | A kind of detection method of license plate based on machine learning | |
| CN102226907B (en) | License plate positioning method and apparatus based on multiple characteristics | |
| CN105913041B (en) | It is a kind of based on the signal lamp recognition methods demarcated in advance | |
| CN101398894B (en) | Automobile license plate automatic recognition method and implementing device thereof | |
| CN105809121A (en) | Multi-characteristic synergic traffic sign detection and identification method | |
| CN103810505B (en) | Vehicles identifications method and system based on multiple layer description | |
| CN103824081B (en) | Method for detecting rapid robustness traffic signs on outdoor bad illumination condition | |
| CN107066972B (en) | Natural scene Method for text detection based on multichannel extremal region | |
| CN102254152A (en) | License plate location method based on color change points and color density | |
| CN104182763A (en) | Plant type identification system based on flower characteristics | |
| CN103366190A (en) | Method for identifying traffic sign | |
| CN103903018A (en) | Method and system for positioning license plate in complex scene | |
| CN102819728A (en) | Traffic sign detection method based on classification template matching | |
| CN103198315A (en) | License plate character segmentation algorithm based on character outline and template matching | |
| CN106529532A (en) | License plate identification system based on integral feature channels and gray projection | |
| CN104182754A (en) | Rural resident point information extraction method based on high-resolution remote-sensing image | |
| CN103390167A (en) | Multi-characteristic layered traffic sign identification method | |
| CN109670515A (en) | Method and system for detecting building change in unmanned aerial vehicle image | |
| CN107220638A (en) | A kind of car plate detection recognition methods based on deep learning convolutional neural networks | |
| CN103093200A (en) | Algorithm for quickly and accurately locating plate number of image | |
| CN102880859A (en) | Method for recognizing number plate | |
| CN105160330A (en) | Vehicle logo recognition method and vehicle logo recognition system | |
| CN103425989A (en) | Vehicle color identification method based on significance analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20100505 |