CN103413135B - A kind of vehicle head lamp bright dark border line detecting method, device and system - Google Patents

Abstract

The present invention relates to technical field of image processing, a kind of bright dark border line detecting method of vehicle head lamp based on image, including: obtain the irradiation image of vehicle head lamp；Rim detection is carried out, it is thus achieved that edge image to irradiating image；In edge image, detection obtains and meets the first line segment aggregate of first direction and meet the second line segment aggregate of second direction respectively, obtains based on the first line segment aggregate and the second line segment aggregate and meets pre-conditioned candidate bright dark side boundary line set；Wherein, direction, first direction correspondence level bright dark side boundary line, the corresponding direction, oblique bright dark side boundary line of second direction；One candidate's bright dark side boundary line is made up of a line segment in the first line segment aggregate and a line segment in the second line segment aggregate；According to vehicle headlight beam feature, the candidate's bright dark side boundary line set obtained determines the bright dark side boundary line of vehicle head lamp.The method that the present invention provides goes for the Outdoor Scene of complexity, and detection sensitivity is high, accuracy is good, strong robustness.

Description

A kind of vehicle head lamp bright dark border line detecting method, device and system

Technical field

The present invention relates to technical field of image processing, particularly relate to a kind of vehicle head lamp based on image Bright dark border line detecting method, device and system.

Background technology

The effect of vehicle head lamp is in the case of night or other low visibility, carries for vehicle driver For illumination, and prevent from headlamp from driver and passerby being caused dazzling the eyes.Headlight beam irradiating angle is too high Or too low all driving can be adversely affected, bring potential safety hazard.General vehicle head lamp is respectively mounted There is the two-filament bulb of instant and near light filament, and during in order to ensure vehicle going at night, Herba Plantaginis is existing bright Bright and illuminate uniformly and can prevent dazzling, distance-light luminous intensity distribution performance and Lighting direction are all had necessarily Requirement.In daily driving process, due to effect vibrated so that headlamp deviation is original Installation site, thus change Lighting direction.Vehicle headlight beam direction of illumination is improper, it has also become Affect one of main hidden danger of Vehicular night safety traffic.Thus it is guaranteed that correct headlamp direction of illumination, Significant to traffic safety.

In prior art, exist and utilize screen or head light checking equipment that headlight beam direction of illumination is carried out The method of detection.When carrying out the detection of headlamp direction of illumination, it will usually according to the light of distance light Yu dipped beam Bundle irradiation feature detects, and wherein, the graphic feature that distance light luminous intensity distribution is formed is symmetric form, Guang Xing center Position is the brightest；The graphic feature that dipped beam luminous intensity distribution is formed is asymmetric, has obvious bright dark side boundary line. Therefore, when detecting lower beam headlamp direction of illumination, the position in bright dark side boundary line is important detection parameter.

Inventor finds during realizing the present invention, and the detection method that prior art exists generally requires Carry out under the specific environments such as measuring station and during stationary vehicle, such as, need when stationary vehicle by projection The position in the image detection headlamp bright dark side boundary line in specific screens.This method cannot be applicable to deposit Outdoor environment in strong interference, it is impossible to detect headlamp bright dark side boundary line in outdoor environment exactly Position.

Summary of the invention

Embodiments provide a kind of bright dark border line detecting method of vehicle head lamp based on image, Device and system, go for existing the vehicle head lamp bright dark side boundary line of the outdoor environment of strong interference Detection, accuracy is high, strong robustness.

Technical scheme is as follows:

First aspect according to embodiments of the present invention, discloses a kind of vehicle head lamp based on image bright secretly Boundary detection, described method includes:

Obtain the irradiation image of vehicle head lamp；

Described irradiation image is carried out rim detection, it is thus achieved that edge image；

In described edge image, detection obtains and meets the first line segment aggregate of first direction and meet respectively Second line segment aggregate of second direction, obtains based on described first line segment aggregate and described second line segment aggregate Meet pre-conditioned candidate bright dark side boundary line set；Wherein, the described first direction bright dark side of correspondence level Direction, boundary line, the corresponding direction, oblique bright dark side boundary line of described second direction；One candidate's bright dark side boundary line by A line segment in described first line segment aggregate and a line segment composition in described second line segment aggregate；

According to vehicle headlight beam feature, before determining vehicle in the candidate's bright dark side boundary line set obtained The bright dark side boundary line of illuminator.

Further, described detection respectively obtains and meets the first line segment aggregate of first direction and meet second Second line segment aggregate in direction includes:

Detection obtains and meets the first line segment aggregate of first direction particularly as follows: each in edge image Marginal point, it is thus achieved that the gradient direction angle of marginal point；Judge gradient direction angle and the first party of marginal point To the difference of corresponding angle whether less than the first predetermined threshold value, it is thus achieved that the first judged result；When described One judged result shows that the difference of the gradient direction angle of the marginal point angle corresponding with first direction is less than the During one predetermined threshold value, preserve described marginal point to the first marginal point set；The marginal point obtained is gathered Class processes, it is thus achieved that corresponding line segment；

Detection meets the second line segment aggregate of second direction particularly as follows: for each edge in edge image Point, it is thus achieved that the gradient direction angle of marginal point；Judge gradient direction angle and the second direction pair of marginal point Whether the difference of the angle answered is less than the second predetermined threshold value, it is thus achieved that the second judged result；When described second sentences Disconnected result shows that the difference of the gradient direction angle of the marginal point angle corresponding with second direction is pre-less than first If during threshold value, preserve described marginal point to the second marginal point set；The marginal point obtained is carried out at cluster Reason, it is thus achieved that corresponding line segment.

Further, the described marginal point to obtaining carries out clustering processing, it is thus achieved that corresponding line segment includes:

Obtain the polar coordinate linear equation parameter ρ corresponding to each marginal point in described marginal point set, press Parameter ρ is sorted by size, and statistics obtains the number of the marginal point with identical parameters ρ, obtains and has local The marginal point set that very big parameter ρ corresponding to marginal point number is corresponding with this parameter ρ；

The marginal point in marginal point set corresponding to each parameter ρ is ranked up；

To the marginal point after sequence, calculate the distance between the point of adjacent two edges, determine according to distance adjacent Whether two edges point belongs to same line segment.

Further, obtain meet default bar based on described first line segment aggregate and described second line segment aggregate One candidate's bright dark side boundary line of part includes:

A line segment is chosen as the first line segment, from described second line-segment sets from described first line segment aggregate Conjunction is chosen a line segment as the second line segment；

Judge the abscissa of terminal of described first line segment whether less than the abscissa of terminal of the second line segment, Obtain the 4th judged result；

Judge the vertical coordinate of terminal of described first line segment whether more than the vertical coordinate of terminal of the second line segment, Obtain the 5th judged result；

Judge that the number of the number of the marginal point of the first line segment and the marginal point of described second line segment is the biggest In the 4th predetermined threshold value, it is thus achieved that the 6th judged result；

When described 4th judged result, described 5th judged result and described 6th judged result are when being, Using described first line segment, described second line segment combination as candidate's bright dark side boundary line.

Further, described according to vehicle headlight beam feature, at the candidate's bright dark side boundary line collection obtained Conjunction determining, bright dark side boundary line includes:

Obtain the response value in each candidate bright dark side boundary line, described response value is ranked up, by response value Big candidate bright dark side boundary line is as the bright dark side boundary line of vehicle head lamp.

Wherein, the response value in each candidate of described acquisition bright dark side boundary line includes:

Obtain in candidate's bright dark side boundary line the meansigma methods of the gradient modulus value of the marginal point of the line segment of first direction with And the meansigma methods of the gradient modulus value of the marginal point of the line segment of second direction；

Obtain the meansigma methods of the gradient modulus value of the marginal point of the line segment of described first direction and described second direction Line segment marginal point gradient modulus value meansigma methods and value as the response in this candidate bright dark side boundary line Value.

Further, in obtaining the combination of candidate's bright dark side boundary line before the response value in each bright dark side boundary line, Described method also includes:

Utilize texture region feature and/or regional luminance feature, further screening candidate's bright dark side boundary line.

Further, described method also includes:

Export positional information and/or the water in this bright dark side boundary line of composition of turning point corresponding to bright dark side boundary line Flat bright dark side boundary line and the positional information in oblique bright dark side boundary line.

Second aspect according to embodiments of the present invention, discloses a kind of vehicle head lamp based on image bright secretly Border line detector, described device includes:

Image acquiring unit, for obtaining the irradiation image of vehicle head lamp；

Edge detection unit, for carrying out rim detection to described irradiation image, it is thus achieved that edge image；

Candidate combinations chooses unit, and in described edge image, detection acquisition meets first party respectively To the first line segment aggregate and meet the second line segment aggregate of second direction, based on described first line segment aggregate Obtain with described second line segment aggregate and meet pre-conditioned candidate bright dark side boundary line set；Wherein, described Direction, first direction correspondence level bright dark side boundary line, the corresponding direction, oblique bright dark side boundary line of described second direction； One candidate's bright dark side boundary line is by a line segment in described first line segment aggregate and described second line segment aggregate In line segment composition；

Determine unit, for according to vehicle headlight beam feature, at the candidate's bright dark side boundary line collection obtained Conjunction determines the bright dark side boundary line of vehicle head lamp.

Further, described candidate combinations is chosen unit and is included that unit is chosen in Line segment detection unit and combination, Described Line segment detection unit includes the first line segment detector unit and the second Line segment detection unit, wherein, described First line segment detector unit includes:

First gradient direction angle acquiring unit, for for each marginal point in edge image, it is thus achieved that The gradient direction angle of marginal point；

First judging unit, for judging the angle that the gradient direction angle of marginal point is corresponding with first direction Difference whether less than the first predetermined threshold value, it is thus achieved that the first judged result；

First storage unit, for when described first judged result show the gradient direction angle of marginal point with When the difference of the angle that first direction is corresponding is less than the first predetermined threshold value, preserve described marginal point to the first limit Edge point set；

Clustering processing unit, for carrying out clustering processing to the marginal point obtained, it is thus achieved that corresponding line segment；

Wherein, described second Line segment detection unit includes:

Second gradient direction angle acquiring unit, for for each marginal point in edge image, it is thus achieved that The gradient direction angle of marginal point；

Second judging unit, for judging the angle that the gradient direction angle of marginal point is corresponding with second direction Difference whether less than the second predetermined threshold value, it is thus achieved that the second judged result；

Second storage unit, for when described second judged result show the gradient direction angle of marginal point with When the difference of the angle that second direction is corresponding is less than the first predetermined threshold value, preserve described marginal point to the second limit Edge point set；

Clustering processing unit, for carrying out clustering processing to the marginal point obtained, it is thus achieved that corresponding line segment.

Further, described clustering processing unit includes:

Marginal point set acquiring unit, for obtaining corresponding to each marginal point in described marginal point set Polar coordinate linear equation parameter ρ, by size to parameter ρ sort, statistics obtain there is identical parameters ρ The number of marginal point, it is corresponding with this parameter ρ that acquisition has the parameter ρ corresponding to local maximum marginal point number Marginal point set；

Sequencing unit, the marginal point in the marginal point set corresponding to each parameter ρ is ranked up；

Line segment processing unit, for the marginal point after sequence, calculating the distance between the point of adjacent two edges, Determine whether adjacent two edges point belongs to same line segment according to distance.

Further, described determine unit specifically for:

Obtain the response value in each candidate bright dark side boundary line, described response value is ranked up, by response value Big candidate bright dark side boundary line is as the bright dark side boundary line of vehicle head lamp.

The third aspect according to embodiments of the present invention, discloses a kind of vehicle head lamp based on image bright secretly Boundary line detecting system, described system includes in-vehicle camera and vehicle head lamp bright dark border line detector, Wherein:

Described in-vehicle camera is for irradiating the irradiation image of vehicle head lamp and bright to described vehicle head lamp Dark border line detector sends the irradiation image of described vehicle head lamp；

Described vehicle head lamp bright dark border line detector is for receiving the vehicle that described in-vehicle camera sends The irradiation image of headlamp；Described irradiation image is carried out rim detection, it is thus achieved that edge image；Described In edge image, detection obtains and meets the first line segment aggregate of first direction and meet second direction respectively Second line segment aggregate, obtains meet default bar based on described first line segment aggregate and described second line segment aggregate Candidate's bright dark side boundary line set of part；Wherein, direction, described first direction correspondence level bright dark side boundary line, The corresponding direction, oblique bright dark side boundary line of described second direction；One described candidate bright dark side boundary line is by described A line segment in one line segment aggregate and a line segment composition in described second line segment aggregate；According to vehicle Headlight beam feature, determines the bright dark side of vehicle head lamp in the candidate's bright dark side boundary line set obtained Boundary line.

What the embodiment of the present invention can reach has the beneficial effect that in the method that the present invention provides, it is thus achieved that car The irradiation image of headlamp, and described irradiation image is carried out rim detection, it is thus achieved that edge image；? In described edge image, detection obtains and meets the first line segment aggregate of first direction and meet second party respectively To the second line segment aggregate, based on described first line segment aggregate and described second line segment aggregate obtain meet pre- If the candidate of condition bright dark side boundary line set；Wherein, described first direction correspondence level bright dark side boundary line side To, the corresponding direction, oblique bright dark side boundary line of described second direction；One described candidate bright dark side boundary line is by institute State a line segment composition in a line segment in the first line segment aggregate and described second line segment aggregate；According to Vehicle headlight beam feature, determines bright dark side boundary line in the candidate's bright dark side boundary line set obtained.? In the method that the present invention provides, owing to dark side boundary line bright to level and oblique bright dark side boundary line are examined respectively Survey and combine, obtain all possible bright dark side boundary line, and utilize headlight beam feature bright from candidate The set of dark side boundary line filters out the bright dark side boundary line of the condition of best suiting as final testing result, effectively Eliminate the impact in the interference edge boundary line caused in outdoor environment due to jamming light source, interference texture, permissible Being applicable to the Outdoor Scene of complexity, detection sensitivity is high, accuracy is good, strong robustness.

Accompanying drawing explanation

In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to reality Execute the required accompanying drawing used in example or description of the prior art to be briefly described, it should be apparent that below, Accompanying drawing in description is only some embodiments described in the present invention, for those of ordinary skill in the art From the point of view of, on the premise of not paying creative work, it is also possible to obtain other accompanying drawing according to these accompanying drawings.

The bright dark border line detecting method of vehicle head lamp based on image first that Fig. 1 provides for the present invention is real Execute example flow chart；

Fig. 2 is filter template schematic diagram；

Fig. 3 is headlamp bright dark side boundary line schematic diagram；

Fig. 4 is that line segment splits schematic diagram；

The bright dark border line detecting method of vehicle head lamp based on image second that Fig. 5 provides for the present invention is real Execute example flow chart；

Vehicle head lamp based on the image bright dark border line detector schematic diagram that Fig. 6 provides for the present invention；

Vehicle head lamp based on the image bright dark side boundary line detecting system schematic diagram that Fig. 7 provides for the present invention.

Detailed description of the invention

Embodiments provide a kind of bright dark border line detecting method of vehicle head lamp based on image, Device and system, go for existing the vehicle head lamp bright dark side boundary line of the outdoor environment of strong interference Detection, accuracy is high, strong robustness.

For the technical scheme making those skilled in the art be more fully understood that in the present invention, below in conjunction with Accompanying drawing in the embodiment of the present invention, clearly and completely retouches the technical scheme in the embodiment of the present invention State, it is clear that described embodiment is only a part of embodiment of the present invention rather than whole enforcement Example.Based on the embodiment in the present invention, those of ordinary skill in the art are not before making creative work Put the every other embodiment obtained, all should belong to the scope of protection of the invention.

First the thought provided the present invention is illustrated.

Before detection during illuminator Lighting direction, particularly during detection dim head lights direction of illumination, logical Often need to detect the position in bright dark side boundary line.This is because, when the light beam of headlamp is radiated at road surface or wall Time on face, the hot spot of certain form can be produced, this hot spot can imaging in in-vehicle camera, formed before automobile Illuminator irradiates image.The hot spot of headlamp is usually expressed as bright dark side boundary line feature in irradiating image, because of The coordinate of the turning point in this position determining bright dark side boundary line and bright dark side boundary line is by vehicle head lamp The important measurement parameter of detection.

In the method that prior art provides, generally require under the specific environments such as measuring station and stationary vehicle The detection of Shi Jinhang headlight beam direction of illumination, before detecting with the image by being projected in specific screens The position in illuminator bright dark side boundary line.This method cannot be applicable to the outdoor environment of complexity.At outdoor environment In, metope or ground etc. are usually present other edges, such as crack, road surface or metope japanning etc., these Dark side bright to headlamp boundary line, edge interferes so that bright dark side boundary line is irradiated in image not at headlamp Continuously, bright dark side boundary line turning point is inconspicuous or there is other limit in addition to headlamp bright dark side boundary line Edge line.Meanwhile, ambient light is crossed bright or bright dark side boundary line the most also can be made to owe clear, thus causes interruption, Have impact on the detection in bright dark side boundary line.Based on this, the method that the present invention provides, according to car headlamp photograph Penetrate the form in the bright dark side of the headlamp in image boundary line, to the segmentation detection of bright dark side boundary line elder generation, then carry out group Close, obtain the headlamp of all candidates bright dark side boundary line in image, and utilize headlight beam feature from time Bright dark side boundary line is selected to wait the bright dark side boundary line filtering out the condition of best suiting in set as final detection knot Really, the shadow in the interference edge boundary line caused in outdoor environment is effectively eliminated due to jamming light source, interference texture Ringing, go for the Outdoor Scene of complexity, detection sensitivity is high, accuracy is good, strong robustness.

See the bright dark border line detecting method of vehicle head lamp based on image that Fig. 1 provides for the present invention One embodiment flow chart.

S101, it is thus achieved that the irradiation image of vehicle head lamp.

When implementing, in-vehicle camera is set at vehicle front, before described in-vehicle camera is used for irradiating vehicle The irradiation image of illuminator, and send described vehicle head lamp to vehicle head lamp bright dark border line detector Irradiation image.Wherein, the shooting direction of in-vehicle camera and the direction of illumination of headlamp are identical.Obtain car The mode of headlamp direction of illumination includes: pre-set the radiation modality of in-vehicle camera, before shooting vehicle Illuminator irradiation image on road surface.The radiation modality of described camera can include on schedule, make a reservation for The mode of mileage, predetermined period or pre-determined number be irradiated one or more.

It will be appreciated by persons skilled in the art that the method that the present invention provides can apply to outdoor environment, And no matter vehicle is in transport condition or halted state is all obtained in that headlamp irradiation figure on road Picture.

S102, carries out rim detection to described irradiation image, it is thus achieved that edge image.

When implementing, the irradiation image obtained is carried out Image Edge-Detection and can use sobel, canny Deng the method at operator extraction edge, edge, this is not defined by the present invention.

In a kind of specific implementation, the present invention carries out Image Edge-Detection the most in the following ways. It will be appreciated by persons skilled in the art that following being merely illustrative, be not regarded as the present invention's Limiting, other implementations that those skilled in the art obtain under not paying creative work belong to this The protection domain of invention.

Specifically, step S102 can be realized by following steps:

S102A, to irradiate image be filtered process, it is thus achieved that horizontal gradient image, vertical gradient image with And the modulus value image of gradient.

When implementing, it is possible to use filter template as shown in Figure 2 is filtered processing, it is thus achieved that water The modulus value image of flat gradient image, vertical gradient image and gradient.

Wherein, described horizontal gradient image I1 obtains in the following manner:

I₁(u, v)=I (u, v) * w^T(u, v) (1)

Wherein, I₁(u, v) represents horizontal gradient image, I (u, v) represents picture point coordinate, and * represents convolution operation, U is picture point horizontal coordinate, and v is picture point vertical coordinate, and w is the matrix of Fig. 2 template representation, w^TIt it is figure The transposition of the matrix of 2 template representation.

Wherein, described vertical gradient image I₂Obtain in the following manner:

I₂(u, v)=I (u, v) * w (u, v) (2)

Wherein, I₂(u, v) represents vertical gradient image, and (u, v) represents picture point coordinate to I, and * represents that convolution is grasped Making, u is picture point horizontal coordinate, and v is picture point vertical coordinate, and w is the matrix of Fig. 2 template representation.

Wherein, the modulus value image Mo of gradient obtains in the following manner:

$\mathrm{Mo}(u,v)=\sqrt{I_1{(u,v)}^2+I_2{(u,v)}^2}---\left(3\right)$

Wherein, (u v) represents the modulus value image of gradient, I to Mo₁(u v) represents horizontal gradient image, I₂(u,v) Represent vertical gradient image.

S102B, according to the modulus value irradiating each pixel of image, it is judged that whether described pixel is marginal point.

When implementing, it is judged that the modulus value of each pixel determines whether described pixel is the one of marginal point Planting possible implementation is: judge that whether the modulus value of pixel is more than predetermined threshold value；When judging pixel Modulus value more than predetermined threshold value time, determine that described pixel is marginal point；When judging the modulus value of pixel not During more than predetermined threshold value, determine that described pixel is non-edge point.For example, it is possible to pre-set threshold value it is 10, when the modulus value of pixel is more than 10, it is determined that described pixel is marginal point, is otherwise non-edge Point.

S102C, will determine as the some composition edge image of marginal point.

Detected that all marginal points, the then figure being made up of all marginal points i.e. form edge by step S102B Testing result figure, i.e. edge image.

S103, in described edge image, detection obtains the first line segment aggregate meeting first direction respectively With meet the second line segment aggregate of second direction, based on described first line segment aggregate and described second line-segment sets Close to obtain and meet pre-conditioned candidate bright dark side boundary line set.

When carrying out the detection of dim head lights direction of illumination, need to detect the position in bright dark side boundary line.Such as Fig. 3 Shown in, for irradiating in image at vehicle head lamp, by level bright dark side boundary line 1, oblique bright dark side boundary line 2, The headlamp bright dark side boundary line schematic diagram that bright dark side boundary line turning point 3 is constituted.

In this step, i.e. will by respectively detection obtain meet first direction the first line segment aggregate and Meet the second line segment aggregate of second direction, the most respectively detection level bright dark side boundary line and oblique bright dark border Line also combines, to obtain all possible bright dark side boundary line.Wherein, described first direction correspondence level is bright Direction, dark side boundary line, the corresponding direction, oblique bright dark side boundary line of described second direction；One described candidate is bright secretly Boundary line is by a line segment in described first line segment aggregate and a line segment in described second line segment aggregate Composition.The set in the corresponding level bright dark side boundary line of set of the first line segment, the set of the second line segment is corresponding tiltedly Set to bright dark side boundary line.Candidate's bright dark side boundary line set is all (right by a level bright dark side boundary line Answer the first line segment) and the bright dark border of candidate that constitutes of an oblique bright dark side boundary line (corresponding second line segment) The set that line is constituted.

When implementing, step S103 specifically can be realized by following steps:

S103A, detection obtains the first line segment aggregate meeting first direction.

Wherein, step S103A can be realized by following steps again:

For each marginal point in edge image, it is thus achieved that the gradient direction angle of marginal point.Judge edge Whether the difference of the gradient direction angle angle corresponding with first direction of point is less than the first predetermined threshold value, obtains Obtain the first judged result；When described first judged result shows gradient direction angle and the first party of marginal point When the difference of corresponding angle is less than the first predetermined threshold value, preserve described marginal point to the first edge point set Close；The marginal point obtained is carried out clustering processing, it is thus achieved that corresponding line segment.

Wherein, the marginal point obtained is carried out clustering processing, it is thus achieved that corresponding line segment includes:

Obtain the polar coordinate linear equation parameter ρ corresponding to each marginal point in described marginal point set, press Parameter ρ is sorted by size, and statistics obtains the number of the marginal point with identical parameters ρ, obtains and has local The marginal point set that very big one or more parameter ρ corresponding to marginal point number are corresponding with parameter ρ；

The marginal point in marginal point set corresponding to each parameter ρ is ranked up；

To the marginal point after sequence, calculate the distance between the point of adjacent two edges, determine according to distance adjacent Whether two edges point belongs to same line segment.

Edge point is carried out first time clustering processing i.e. by obtaining each edge in described marginal point set Polar coordinate linear equation parameter ρ corresponding to Dian, sorts to parameter ρ by size, and statistics acquisition has identical The number of the marginal point of parameter ρ, obtains and has the parameter ρ corresponding to local maximum marginal point number and this ginseng Count marginal point set corresponding for ρ has the beneficial effects that the Clutter edge point that can get rid of in background in a large number, has Effect improves the accuracy of detection.

It is implemented as: for each marginal point, (u v), if its gradient direction | θ-φ |≤θ _ thre, then leads to Cross the parameter ρ that formula (4) calculates the polar coordinate linear equation of its correspondence:

ρ=ucos (θ)+vsin (θ) (4)

Wherein, φ is the orientation angle specified, and the angle for first direction could be arranged to 90 degree, and θ _ thre is Gradient direction error threshold, θ represents linear gradient orientation angle.

$\mathrm{\θ}=\arg \tan \frac{I_2(u,v)}{I_1(u,v)}---\left(5\right)$

Sorting parameter ρ by size, statistics obtains the number of the marginal point with identical parameters ρ, obtains tool There is the marginal point set that the parameter ρ corresponding to local maximum marginal point number is corresponding with this parameter ρ；Obtaining After the marginal point set that the parameter ρ of maximum point is corresponding, the marginal point set to each acquisition is carried out at cluster Reason, it is thus achieved that some first line segments.Concrete, the marginal point in the marginal point set S detected is carried out Sequence: the name placement of pressing in set S is obtained ordered set S ', sequence need to meet: for arbitrarily A bit, if it has a left side (right) adjoint point, then it is to the Euclidean distance of left (right) adjoint point, is not more than it and arrives left (right) The Euclidean distance of other point of side.The little point of abscissa is in left side, if abscissa is identical, then vertical coordinate is big Point in left side.

After completing sequence, to the marginal point after sequence, calculate the distance between the point of adjacent two edges, according to Distance determines whether adjacent two edges point belongs to same line segment.Concrete, it is thus achieved that between the point of adjacent two edges Distance, it is judged that whether the distance between the point of adjacent two edges more than the 3rd predetermined threshold value, it is thus achieved that the 3rd sentences Disconnected result；When described 3rd judged result shows that the distance between the point of adjacent two edges presets threshold more than the 3rd During value, then carry out line segment dividing processing；When described 3rd judged result shows between the point of adjacent two edges When distance is less than three predetermined threshold value, it is determined that adjacent two edges point belongs to same line segment.By searching phase The distance of adjacent marginal point carries out line segment segmentation more than the position of the 3rd threshold value D_thre and is advantageous in that and can solve The line segment discontinuous problem that certainly boundary line is caused by environmental disturbances.Segmentation schematic diagram is as shown in Figure 4.As Shown in Fig. 4, Fig. 4 has 21 marginal points, is divided into two line segments through line segment segmentation.The left side 6 Individual marginal point constitutes line segment 11, and 15, the right marginal point constitutes line segment 22.Point A is 11, line segment Point, some F is line segment 11 terminal.Point G is line segment 22 starting point, and some U is line segment 22 terminal.Pass through Calculating, distance D between some F and some G, more than the 3rd threshold value D_thre, carries out line segment the most from there Segmentation, it is believed that some F is not belonging to same line segment with some G.

S103B, detection has the second line segment of second direction.

When implementing, step S103B can be realized by following steps again:

For each marginal point in edge image, it is thus achieved that the gradient direction angle of marginal point.Judge edge Whether the difference of the gradient direction angle angle corresponding with second direction of point is less than the second predetermined threshold value, obtains Obtain the second judged result；When described second judged result shows gradient direction angle and the second party of marginal point When the difference of corresponding angle is less than the first predetermined threshold value, preserve described marginal point to the second edge point set Close；The marginal point obtained is carried out clustering processing, it is thus achieved that corresponding line segment.

Concrete, the marginal point obtained is carried out clustering processing, it is thus achieved that corresponding line segment includes:

Obtain the polar coordinate linear equation parameter ρ corresponding to each marginal point in described marginal point set, press Parameter ρ is sorted by size, and statistics obtains the number of the marginal point with identical parameters ρ, obtains and has local The marginal point set that very big parameter ρ corresponding to marginal point number is corresponding with this parameter ρ；

The marginal point in marginal point set corresponding to each parameter ρ is ranked up；

To the marginal point after sequence, calculate the distance between the point of adjacent two edges, determine according to distance adjacent Whether two edges point belongs to same line segment.

The process of implementing is referred to step S103A and carries out, wherein, it is intended that the angle of second direction Could be arranged to 75 degree.

S103C, meets pre-conditioned based on described first line segment aggregate and described second line segment aggregate acquisition Candidate's bright dark side boundary line.

(oblique to the first line segment (level bright dark border line segment) detected in step S102 and the second line segment To bright dark border line segment) it is combined, obtain complete bright dark side boundary line.Each combination is by a level Bright dark border line segment and an oblique bright dark border line segment are constituted, and can set combination bar according to practical situation Part.Such as require the end point (u of level bright dark border line segment in image₁,v₁) occur in the initial of oblique line segment Point (u₂,v₂) lower left, and the some number in level bright dark side boundary line and oblique bright dark side boundary line is at reasonable model In enclosing.When implementing, meet pre-based on described first line segment aggregate and described second line segment aggregate acquisition If condition candidate's bright dark side boundary line includes: choose a line segment from described first line segment aggregate and make It is the first line segment, from described second line segment aggregate, chooses a line segment as the second line segment；Judge described Whether the abscissa of the terminal of the first line segment is less than the abscissa of the terminal of the second line segment, it is thus achieved that the 4th judges Result；Judge the vertical coordinate of terminal of described first line segment whether more than the vertical coordinate of terminal of the second line segment, Obtain the 5th judged result；Judge number and the edge of described second line segment of marginal point on the first line segment Whether the number of point is all higher than the 4th predetermined threshold value, it is thus achieved that the 6th judged result；When the described 4th judges knot Really, described 5th judged result and described 6th judged result are when being, by described first line segment, institute State the combination of the second line segment as candidate's bright dark side boundary line.Wherein, the 4th predetermined threshold value can be in advance Set.Repeat the above steps, i.e. can obtain multiple candidate bright dark side boundary line.

S104, according to vehicle headlight beam feature, determines in the candidate's bright dark side boundary line set obtained Bright dark side boundary line.

The purpose of step S104 is from the some candidates bright dark side boundary line obtained to select to best suit front photograph The bright dark side boundary line of lamp feature.When implementing, step S104 can have multiple implementation, then walk Rapid S104 specifically may include that

Obtain the response value of each candidate bright dark side boundary line set, described response value is ranked up, will response Candidate's bright dark side boundary line of value maximum is as the bright dark side boundary line of vehicle head lamp.

Wherein, the response value in each candidate of described acquisition bright dark side boundary line includes:

Obtain in candidate's bright dark side boundary line the meansigma methods of the gradient modulus value of the marginal point of the line segment of first direction with And the meansigma methods of the gradient modulus value of the marginal point of the line segment of second direction；

Obtain the meansigma methods of the gradient modulus value of the marginal point of the line segment of described first direction and described second direction Line segment marginal point gradient modulus value meansigma methods and value as the response in this candidate bright dark side boundary line Value.

In another implementation of the present invention, each bright dark side boundary line in obtaining the combination of candidate's bright dark side boundary line Response value before, described method also includes:

Utilize texture region feature and/or regional luminance feature, further screening candidate's bright dark side boundary line.

It is to say, described method also includes utilizing texture region feature, in the bright dark border of candidate obtained Line set is screened candidate's bright dark side boundary line；And/or

Utilize regional luminance feature, the candidate's bright dark side boundary line set obtained is screened the bright dark border of candidate Line；And/or

Below specific implementation is introduced.It will be appreciated by persons skilled in the art that can be only Response value feature is used to be ranked up determining bright dark side to the response value of all bright dark side boundary lines candidate combinations The final position in boundary line.Can also first use textural characteristics, determine the position candidate in bright dark side boundary line, so Rear recycling response value feature determines the final position in bright dark side boundary line.Can also first use regional luminance special Levying the position candidate determining bright dark side boundary line, then recycling response value feature determines that bright dark side boundary line is Final position is put.Certainly, it will be appreciated by persons skilled in the art that can also utilize simultaneously textural characteristics and Regional luminance feature determines the position candidate in bright dark side boundary line, then recycling response value feature determine bright secretly The final position of boundary line.Being merely illustrative additionally, following, those skilled in the art are not paying Other implementations obtained under creative work belong to protection scope of the present invention.

Wherein, utilize texture region feature, the candidate's bright dark side boundary line set obtained is screened candidate bright Dark side boundary line may include that in the candidate's bright dark side boundary line set obtained, it is thus achieved that bright dark side boundary line is corresponding Texture region；Judge that whether the entropy of texture region is less than the 5th predetermined threshold value；When described texture region Entropy less than five predetermined threshold value time, determine that described texture region is possible headlamp region, by institute State bright dark side boundary line corresponding to texture region as candidate's bright dark side boundary line.When implementing, headlamp district The texture in territory should meet certain condition, and in image based on road surface, upper beam spot region belongs to weak Texture region, if texture strength is too big near bright dark side boundary line, then may be barrier region, Ying Yupai Remove.In practical operation: texture region scope corresponding to bright dark side boundary line can external by bright dark side boundary line Rectangle determines.The power of horizontal amount texture can be judged by the entropy in region.Such as, it is judged that texture region Entropy whether less than the 5th predetermined threshold value, when the entropy of described texture region is less than five predetermined threshold value, Determine that described texture region is the potential region of headlamp, bright dark side boundary line corresponding for described texture region is made For candidate's bright dark side boundary line.

Wherein, utilize regional luminance feature, screen follow-up bright in the candidate's bright dark side boundary line set obtained Dark side boundary line may include that and obtains the brightness of the upper area of the first line segment in candidate's bright dark side boundary line set The brightness value of the lower zone of value and the first line segment；Obtain the second line segment in candidate's bright dark side boundary line set The brightness value of left field and the brightness value of the right side area of the second line segment；Judge the lower zone of the first line segment Whether the brightness value in territory is more than the 6th predetermined threshold value, it is thus achieved that the 7th judged result；Judge the right side of the second line segment Whether the brightness value of side region is more than the 7th predetermined threshold value, it is thus achieved that the 8th judged result；Judge the first line segment The brightness value of lower zone and the first line segment upper area brightness value between difference whether more than the Eight predetermined threshold value, it is thus achieved that the 9th judged result；Judge the brightness value and second of the right side area of the second line segment Whether the difference between the brightness value in the region, left of line segment is more than the 9th predetermined threshold value, it is thus achieved that the tenth judges Result；When described 7th judged result, described 8th judged result, described 9th judged result, described Tenth judged result is when being, using the combination of described first line segment and the second line segment as the bright dark side of candidate Boundary line.When implementing, about headlamp level lower regions, bright dark side boundary line and oblique bright dark side boundary line Image intensity contrast's degree in region should be in the reasonable scope.Assume under the first line segment i.e. level bright dark side boundary line Side region E₁Interior average luminance of pixels value is I (E₁), the first line segment i.e. level bright dark side boundary line upper area E₂Interior average luminance of pixels value is I (E₂), the second line segment the most oblique bright dark side boundary line right side area E₃In Average luminance of pixels value is I (E₃), the second line segment the most oblique bright dark side boundary line left field E₄Interior pixel is put down All brightness values are I (E₄).Region E₁, E₂, E₃, E₄Form and dimension can set according to practical situation. Headlamp bright dark side boundary line should meet following condition:

$\left\{\begin{array}{c}I\left(E_1\right)>{\mathrm{IE}}_1\_\mathrm{thre}\\ I\left(E_3\right)>{\mathrm{IE}}_3\_\mathrm{thre}\\ I\left(E_1\right)-I\left(E_2\right)>{\mathrm{IE}}_{12}\_\mathrm{thre}\\ I\left(E_3\right)-I\left(E_4\right)>{\mathrm{IE}}_{34}\_\mathrm{thre}\end{array}---\left(6\right)\right.$

Wherein, IE₁_ thre is the 6th predetermined threshold value, IE₃_ thre is the 7th predetermined threshold value, IE₁₂_ thre is Eight predetermined threshold value, IE₃₄_ thre is the 9th predetermined threshold value, all can set according to practical situation.

Wherein, described according to vehicle headlight beam feature, at the candidate's bright dark side boundary line Candidate Set obtained Conjunction determining, bright dark side boundary line includes: obtain the response value in each candidate bright dark side boundary line, to described response value It is ranked up, using candidate's bright dark side boundary line maximum for response value as the bright dark side boundary line of vehicle head lamp. When implementing: obtain the ladder of the marginal point of the first line segment of first direction in candidate's bright dark side boundary line set The meansigma methods of the gradient modulus value of the marginal point of the meansigma methods of degree modulus value and the second line segment of second direction；Obtain Obtain the meansigma methods of the gradient modulus value of the marginal point of described first line segment and the ladder of the marginal point of described second line segment The meansigma methods of degree modulus value and value as the response value in this candidate bright dark side boundary line；By maximum for response value bright The position of dark side boundary line combination correspondence is as the final position in bright dark side boundary line.

Wherein, the gradient modulus value of the marginal point of the first line segment obtains in the following manner:

$I\left(M_H\right)=\sqrt{I_1{(u,v)}^2+I_2{(u,v)}^2}---\left(7\right)$

Wherein, I (M_H) represent the modulus value of gradient of marginal point, I₁(u v) represents the marginal point of the first line segment Horizontal gradient image, I₂(u v) represents the vertical gradient image of the marginal point of the first line segment.

Obtain the modulus value of all marginal points in the first line segment i.e. level bright dark side boundary line, then calculate average Value, has i.e. obtained the meansigma methods of the gradient modulus value of the marginal point of the first line segment.

Wherein, the Grad modulus value of the marginal point of the second line segment obtains in the following manner:

$I\left(M_S\right)=\sqrt{I_1{(u,v)}^2+I_2{(u,v)}^2}---\left(8\right)$

Wherein, I (M_S) represent the modulus value of gradient of marginal point, I₁(u v) represents the marginal point of the second line segment Horizontal gradient image, I₂(u v) represents the vertical gradient image of the marginal point of the second line segment.

Wherein, response value I (M) obtains in the following manner:

I(M)=I(M_H)+I(M_S) (9)

Response value is that point is met the value that bright dark side boundary line characteristic length is described.When modulus value is the biggest, say It is the strongest that bright this lights dark change, and this just meets the characteristics of image in bright dark side boundary line, therefore can be by modulus value A kind of form as response value.Other can describe certain point and belong in bright dark side boundary line the retouching of probability of point State, can serve as response value.

In this embodiment of the present invention, position corresponding for bright dark side boundary line combination maximum for response value is made Final position for bright dark side boundary line.

If it should be noted that screened further first with textural characteristics or regional luminance feature Candidate's bright dark side boundary line, then the bright dark side boundary line calculating response value is the bright dark border of described candidate filtered out Line.

Described method also includes: export the positional information in bright dark side boundary line.

Wherein, the positional information in bright dark side boundary line includes the positional information of turning point corresponding to bright dark side boundary line And/or form level bright dark side boundary line and the positional information in oblique bright dark side boundary line in this bright dark side boundary line.Institute State the positional information positional information and described by described first line segment of turning point corresponding to bright dark side boundary line The positional information of two line segments determines.Concrete, the positional information of the turning point that bright dark side boundary line is corresponding is concrete For level bright dark side boundary line and the intersecting point coordinate in oblique bright dark side boundary line.

Further, the method that the present invention provides may include that

Utilize positional information and/or this bright dark side boundary line of composition of turning point corresponding to described bright dark side boundary line Level bright dark side boundary line and the irradiating angle of positional information calculation headlamp in oblique bright dark side boundary line.

In the first embodiment of the invention, in the method that the present invention provides, due to meeting first direction Level bright dark side boundary line and oblique bright dark side boundary line detect respectively, obtain all possible bright secretly The candidate combinations of boundary line, and utilize headlight beam feature to filter out from candidate's bright dark side boundary line set Best suit the bright dark side boundary line of condition as final testing result, effectively eliminate in outdoor environment due to The impact in the interference edge boundary line that jamming light source, interference texture cause, goes for the Outdoor Scene of complexity, Detection sensitivity is high, accuracy is good, strong robustness.

Especially, in complex scene, there is jamming light source, interference texture etc. causes other to be similar to There is the headlamp bright dark side boundary line caused and detect inaccurate asking in the boundary line in headlamp bright dark side boundary line Topic, the present invention obtains all possible bright dark side boundary line by segmentation detection and combines, and according to headlamp Characteristics of image, therefrom screening best suits the boundary line of condition, as headlamp bright dark side boundary line, effective gram Having taken interference, detection sensitivity is high, accuracy is good, strong robustness.

Especially, in complex scene, headlamp bright dark side boundary line can be caused due to interference texture etc. Headlamp bright dark side boundary line that is discontinuous and that cause detects inaccurate problem, in the present invention, each bright secretly The testing result of boundary line, is all to obtain by the way of combination, situation separated in allowing;And again When having bright dark side boundary line marginal point to obtain bright dark border line segment, it is provided with threshold value D_thre, and judges adjacent Whether the distance of point is defined segmentation more than setting threshold value.Therefore, for every section of line segment, when When edge is interrupted less than D_thre, still can be linked to be same line segment, will not disconnect.Therefore, limit is worked as When boundary line is discontinuous, still can accurately detect, strong robustness.

See Fig. 5, the bright dark border line detecting method of vehicle head lamp based on image the provided for the present invention Two embodiment flow charts.

In this specific implementation, determine the final of bright dark side boundary line utilizing headlight beam feature During position, the present invention considers headlight beam feature i.e. texture region feature and/or brightness is carried out Screening, and in the candidate combinations filtered out, using position corresponding for combination maximum for response value as bright secretly The final position of boundary line.

S501, it is thus achieved that the irradiation image of vehicle head lamp.

Specific implementation refer to the implementation of step S101.

S502, carries out rim detection to described irradiation image, it is thus achieved that edge image.

Specific implementation refer to the implementation of step S102.

S503, in described edge image, detection obtains the first line segment aggregate meeting first direction respectively With meet the second line segment aggregate of second direction, based on described first line segment aggregate and described second line-segment sets Close to obtain and meet pre-conditioned candidate bright dark side boundary line set.

Specific implementation refer to the implementation of step S103.

S504, utilizes texture region feature to screen candidate in the candidate's bright dark side boundary line set obtained bright secretly Boundary line.

When implementing, utilize texture region feature, screen in the candidate's bright dark side boundary line set obtained Candidate's bright dark side boundary line includes: in the candidate's bright dark side boundary line set obtained, it is thus achieved that bright dark side boundary line pair The texture region answered；Judge that whether the entropy of texture region is less than the 5th predetermined threshold value；When described texture area When the entropy in territory is less than five predetermined threshold value, determine that described texture region is headlamp candidate region, by institute State bright dark side boundary line corresponding to texture region as candidate's bright dark side boundary line.

S505, utilizes regional luminance feature, screens candidate bright in the candidate's bright dark side boundary line set obtained Dark side boundary line.

When implementing, utilize regional luminance feature, screen in the candidate's bright dark side boundary line set obtained Candidate's bright dark side boundary line includes: in acquisition candidate's bright dark side boundary line set, the upper area of the first line segment is bright The brightness value of the lower zone of angle value and the first line segment；Obtain the second line segment in candidate's bright dark side boundary line set The brightness value of left field and the brightness value of right side area of the second line segment；Judge the lower section of the first line segment Whether the brightness value in region is more than the 6th predetermined threshold value, it is thus achieved that the 7th judged result；Judge the second line segment Whether the brightness value of right side area is more than the 7th predetermined threshold value, it is thus achieved that the 8th judged result；Judge First Line Whether the difference between the brightness value of the brightness value of the lower zone of section and the upper area of the first line segment is more than 8th predetermined threshold value, it is thus achieved that the 9th judged result；Judge the brightness value and of the right side area of the second line segment Whether the difference between the brightness value in the region, left of two line segments is more than the 9th predetermined threshold value, it is thus achieved that the tenth sentences Disconnected result；When described 7th judged result, described 8th judged result, described 9th judged result, institute State the tenth judged result and be when being, using bright secretly as candidate for the combination of described first line segment and the second line segment Boundary line.

It will be appreciated by persons skilled in the art that step S504 and step S505 do not have the priority of certainty Execution sequence, can be performed in parallel, it is also possible to perform reversedly.Certainly, step S504 and step S505 One can also be selected be suitable for.

S506, in the candidate filtered out bright dark side boundary line, utilizes response value feature, determines bright dark border Line.

In all candidate's bright dark side boundary lines meeting above-mentioned condition, headlamp bright dark side boundary line is for having The bright dark side boundary line of big response value.Such as can add up the gradient of point in the level bright dark side boundary line of residue combination Meansigma methods I (the M of modulus value_H), the meansigma methods I (M of the gradient modulus value of point in oblique bright dark side boundary line_S).Calculate Response value I (M):

I (M)=I (M_H)+I(M_S) (10)

Then, take and there is position corresponding to the combination of maximum I (M) value as headlamp bright dark side boundary line Final position is put.

S507, the positional information in the described bright dark side boundary line that output determines.

Vehicle head lamp based on the image bright dark border line detector schematic diagram that Fig. 6 provides for the present invention.

A kind of bright dark border line detector of vehicle head lamp based on image, described device includes:

Image acquiring unit 601, for obtaining the irradiation image of vehicle head lamp.

Edge detection unit 602, for carrying out rim detection to described irradiation image, it is thus achieved that edge image.

Candidate combinations chooses unit 603, and in described edge image, detection acquisition meets first respectively First line segment aggregate in direction and the second line segment aggregate meeting second direction, based on described first line-segment sets Close and described second line segment aggregate obtains and meets pre-conditioned candidate bright dark side boundary line set；Wherein, institute State direction, first direction correspondence level bright dark side boundary line, the corresponding oblique bright dark side boundary line side of described second direction To；One described candidate bright dark side boundary line is by the line segment and described second in described first line segment aggregate A line segment composition in line segment aggregate.

Determine unit 604, for according to headlight beam feature, in the candidate's bright dark side boundary line set obtained The middle bright dark side boundary line determining vehicle head lamp.

Further, described device also includes:

Output unit, for exporting the positional information in bright dark side boundary line.

Further, described candidate combinations is chosen unit and is included that unit is chosen in Line segment detection unit and combination, Wherein, described Line segment detection unit the first line segment detector unit and the second Line segment detection unit, wherein, the One Line segment detection unit includes:

First gradient direction angle acquiring unit, for for each marginal point in edge image, it is thus achieved that The gradient direction angle of marginal point；

First judging unit, for judging the angle that the gradient direction angle of marginal point is corresponding with first direction Difference whether less than the first predetermined threshold value, it is thus achieved that the first judged result；

First storage unit, for when described first judged result show the gradient direction angle of marginal point with When the difference of the angle that first direction is corresponding is less than the first predetermined threshold value, preserve described marginal point to the first limit Edge point set；

Clustering processing unit, for carrying out clustering processing to the marginal point obtained, it is thus achieved that corresponding line segment.

Further, described candidate combinations is chosen unit and is included that unit is chosen in Line segment detection unit and combination, Wherein, described Line segment detection unit the first line segment detector unit and the second Line segment detection unit, wherein, the Two line segment detector units include:

Second gradient direction angle acquiring unit, for for each marginal point in edge image, it is thus achieved that The gradient direction angle of marginal point；

Second judging unit, for judging the angle that the gradient direction angle of marginal point is corresponding with second direction Difference whether less than the second predetermined threshold value, it is thus achieved that the second judged result；

Second storage unit, for when described second judged result show the gradient direction angle of marginal point with When the difference of the angle that second direction is corresponding is less than the first predetermined threshold value, preserve described marginal point to the second limit Edge point set；

Clustering processing unit, for carrying out clustering processing to the marginal point obtained, it is thus achieved that corresponding line segment.

Further, described clustering processing unit includes:

Marginal point set acquiring unit, for obtaining corresponding to each marginal point in described marginal point set Polar coordinate linear equation parameter ρ, by size to parameter ρ sort, statistics obtain there is identical parameters ρ The number of marginal point, it is corresponding with this parameter ρ that acquisition has the parameter ρ corresponding to local maximum marginal point number Marginal point set；

Sequencing unit, the marginal point in the marginal point set corresponding to each parameter ρ is ranked up；

Line segment processing unit, for the marginal point after sequence, calculating the distance between the point of adjacent two edges, Determine whether adjacent two edges point belongs to same line segment according to distance.

Further, described candidate combinations is chosen unit and is included that unit is chosen in Line segment detection unit and combination, Wherein, described combination is chosen unit and is included:

Choose unit, for choosing a line segment from described first line segment aggregate as the first line segment, from Described second line segment aggregate is chosen a line segment as the second line segment；

4th judging unit, for judging that whether the abscissa of the terminal of described first line segment is less than the second line The abscissa of the terminal of section, it is thus achieved that the 4th judged result；

5th judging unit, for judging that whether the vertical coordinate of the terminal of described first line segment is more than the second line The vertical coordinate of the terminal of section, it is thus achieved that the 5th judged result；

6th judging unit, for judging the number of marginal point on the first line segment and described second line segment Whether the number of marginal point is all higher than the 4th predetermined threshold value, it is thus achieved that the 6th judged result；

Assembled unit, for sentencing when described 4th judged result, described 5th judged result and the described 6th Disconnected result is when being, using described first line segment, described second line segment combination as the bright dark border of candidate Line.

Further, described determine unit specifically for:

Obtain the response value in each bright dark side boundary line, described response value is ranked up, by response value maximum Candidate's bright dark side boundary line is as the bright dark side boundary line of vehicle head lamp.

Further, described determine that unit is specifically for obtaining first direction in candidate's bright dark side boundary line set The meansigma methods of gradient modulus value of marginal point of the first line segment and the marginal point of the second line segment of second direction The meansigma methods of gradient modulus value；Obtain the gradient modulus value of the marginal point of the first line segment of described first direction The meansigma methods of the gradient modulus value of the marginal point of the second line segment of meansigma methods and described second direction and value conduct The response value in bright dark side boundary line.

Further, described device also includes:

First determines unit, is used for utilizing texture region feature, in the candidate's bright dark side boundary line set obtained In determine candidate's bright dark side boundary line；And/or

Second determines unit, is used for utilizing regional luminance feature, in the candidate's bright dark side boundary line set obtained In determine candidate's bright dark side boundary line.The most described determine unit, be used for utilizing response value feature, obtain Candidate's bright dark side determines bright dark side boundary line in boundary line.

Further, described first determines that unit includes:

Texture region obtains unit, in the candidate's bright dark side boundary line set obtained, it is thus achieved that bright dark side The texture region that boundary line is corresponding；

6th judging unit, for judging that whether the entropy of texture region is less than the 5th predetermined threshold value；

Primary importance determines unit, is used for when the entropy of described texture region is less than five predetermined threshold value, Determine that described texture region is headlamp candidate region, bright dark side boundary line corresponding for described texture region is made For candidate's bright dark side boundary line.

Further, described second determines that unit includes:

Brightness value obtains unit, for obtaining the upper area of the first line segment in candidate's bright dark side boundary line set Brightness value and the brightness value of lower zone of the first line segment；Obtain in candidate's bright dark side boundary line set second The brightness value of the left field of line segment and the brightness value of the right side area of the second line segment；

7th judging unit, the most pre-more than the 6th for judging the brightness value of the lower zone of the first line segment If threshold value, it is thus achieved that the 7th judged result；

8th judging unit, the most pre-more than the 7th for judging the brightness value of the right side area of the second line segment If threshold value, it is thus achieved that the 8th judged result；

9th judging unit, for judging the brightness value of lower zone of the first line segment and the first line segment Whether the difference between the brightness value in region, side is more than the 8th predetermined threshold value, it is thus achieved that the 9th judged result；

Tenth judging unit, for judging brightness value and the left side for the second line segment of the right side area of the second line segment Whether the difference between the brightness value in region, side is more than the 9th predetermined threshold value, it is thus achieved that the tenth judged result；

The second position determines unit, for when described 7th judged result, described 8th judged result, institute State the 9th judged result, described tenth judged result is when being, by described first line segment and the second line segment Combination as candidate's bright dark side boundary line.

Further, described determine that unit includes:

Modulus value acquiring unit, for obtaining the ladder of the marginal point of the first line segment in candidate's bright dark side boundary line set The meansigma methods of the gradient modulus value of the meansigma methods of degree modulus value and the marginal point of the second line segment；

Response value acquiring unit, for obtaining the meansigma methods of the gradient modulus value of the marginal point of described first line segment With the meansigma methods of the gradient modulus value of the marginal point of described second line segment and value as the response in bright dark side boundary line Value；

3rd position determination unit, for the bright dark side boundary line that response value is maximum as vehicle head lamp Bright dark side boundary line.

Further, described output unit specifically for:

Export positional information and/or the water in this bright dark side boundary line of composition of turning point corresponding to bright dark side boundary line Flat bright dark side boundary line and the positional information in oblique bright dark side boundary line；The turning point that described bright dark side boundary line is corresponding Positional information determined by the positional information of described first line segment and the positional information of described second line segment；

The most described device also includes:

Irradiating angle computing unit, is used for the positional information of the turning point utilizing described bright dark side boundary line corresponding And/or form level bright dark side boundary line and the positional information meter in oblique bright dark side boundary line in this bright dark side boundary line Calculate the irradiating angle of headlamp.

Vehicle head lamp based on the image bright dark side boundary line detecting system schematic diagram that Fig. 7 provides for the present invention.

A kind of vehicle head lamp bright dark side boundary line based on image detecting system, described system includes vehicle-mounted phase Machine and vehicle head lamp bright dark border line detector, wherein:

Described in-vehicle camera 701 is for irradiating the irradiation image of vehicle head lamp, and shines before described vehicle Lamp bright dark border line detector sends the irradiation image of described vehicle head lamp；

Described vehicle head lamp bright dark border line detector 702 is for receiving what described in-vehicle camera sent The irradiation image of vehicle head lamp；Described irradiation image is carried out Image Edge-Detection, it is thus achieved that edge image； In described edge image, detection obtains and meets the first line segment aggregate of first direction and meet second respectively Second line segment aggregate in direction, meets default based on described first line segment aggregate and described second line segment aggregate Candidate's bright dark side boundary line set of condition；Wherein, direction, described first direction correspondence level bright dark side boundary line, The corresponding direction, oblique bright dark side boundary line of described second direction；In described candidate bright dark side boundary line set one Candidate's bright dark side boundary line is by a line segment in described first line segment aggregate and described second line segment aggregate One line segment composition；According to vehicle headlight beam feature, in the candidate's bright dark side boundary line set obtained Determine the bright dark side boundary line of vehicle head lamp.

It will be appreciated by persons skilled in the art that vehicle head lamp of the present invention bright dark side boundary line detecting system Including vehicle head lamp bright dark border line detector can be realized by the device shown in Fig. 6.

It should be noted that in this article, the relational terms of such as first and second or the like is used merely to One entity or operation are separated with another entity or operating space, and not necessarily requires or imply Relation or the order of any this reality is there is between these entities or operation.And, term " includes ", " comprise " or its any other variant is intended to comprising of nonexcludability, so that include a series of wanting Process, method, article or the equipment of element not only include those key elements, but also include the most clearly arranging Other key elements gone out, or also include want intrinsic for this process, method, article or equipment Element.In the case of there is no more restriction, statement " including ... " key element limited, it is not excluded that Other identical element is there is also in including the process of described key element, method, article or equipment.

The present invention can described in the general context of computer executable instructions, Such as program module.Usually, program module includes performing particular task or realizing specific abstract data class The routine of type, program, object, assembly, data structure etc..Can also be in a distributed computing environment Put into practice the present invention, in these distributed computing environment, by by communication network connected remotely Reason equipment performs task.In a distributed computing environment, program module may be located at and includes storage device In interior local and remote computer-readable storage medium.

The above is only the detailed description of the invention of the present invention, it is noted that general for the art For logical technical staff, under the premise without departing from the principles of the invention, it is also possible to make some improvement and profit Decorations, these improvements and modifications also should be regarded as protection scope of the present invention.

Claims (12)

1. the bright dark border line detecting method of vehicle head lamp based on image, it is characterised in that described Method includes:

Obtain the irradiation image of vehicle head lamp；

Described irradiation image is carried out rim detection, it is thus achieved that edge image；

In described edge image, detection obtains and meets the first line segment aggregate of first direction and meet respectively Second line segment aggregate of second direction, obtains based on described first line segment aggregate and described second line segment aggregate Meet pre-conditioned candidate bright dark side boundary line set；Wherein, the described first direction bright dark side of correspondence level Direction, boundary line, the corresponding direction, oblique bright dark side boundary line of described second direction；One candidate's bright dark side boundary line by A line segment in described first line segment aggregate and a line segment composition in described second line segment aggregate；

According to vehicle headlight beam feature, before determining vehicle in the candidate's bright dark side boundary line set obtained The bright dark side boundary line of illuminator.

Method the most according to claim 1, it is characterised in that the described acquisition of detection respectively meets the First line segment aggregate in one direction and the second line segment aggregate meeting second direction include:

Detection obtains and meets the first line segment aggregate of first direction particularly as follows: each in edge image Marginal point, it is thus achieved that the gradient direction angle of marginal point；Judge gradient direction angle and the first party of marginal point To the difference of corresponding angle whether less than the first predetermined threshold value, it is thus achieved that the first judged result；When described One judged result shows that the difference of the gradient direction angle of the marginal point angle corresponding with first direction is less than the During one predetermined threshold value, preserve described marginal point to the first marginal point set；The marginal point obtained is gathered Class processes, it is thus achieved that corresponding line segment；

Detection meets the second line segment aggregate of second direction particularly as follows: for each edge in edge image Point, it is thus achieved that the gradient direction angle of marginal point；Judge gradient direction angle and the second direction pair of marginal point Whether the difference of the angle answered is less than the second predetermined threshold value, it is thus achieved that the second judged result；When described second sentences Disconnected result shows that the difference of the gradient direction angle of the marginal point angle corresponding with second direction is pre-less than first If during threshold value, preserve described marginal point to the second marginal point set；The marginal point obtained is carried out at cluster Reason, it is thus achieved that corresponding line segment.

Method the most according to claim 2, it is characterised in that the described marginal point to obtaining is carried out Clustering processing, it is thus achieved that corresponding line segment includes:

Obtain the polar coordinate linear equation parameter ρ corresponding to each marginal point in described marginal point set, press Parameter ρ is sorted by size, and statistics obtains the number of the marginal point with identical parameters ρ, obtains and has local The marginal point set that very big parameter ρ corresponding to marginal point number is corresponding with this parameter ρ；

The marginal point in marginal point set corresponding to each parameter ρ is ranked up；

To the marginal point after sequence, calculate the distance between the point of adjacent two edges, determine according to distance adjacent Whether two edges point belongs to same line segment.

Method the most according to claim 1, it is characterised in that based on described first line segment aggregate and Described second line segment aggregate acquisition meets a pre-conditioned candidate bright dark side boundary line and includes:

A line segment is chosen as the first line segment, from described second line-segment sets from described first line segment aggregate Conjunction is chosen a line segment as the second line segment；

Judge the abscissa of terminal of described first line segment whether less than the abscissa of terminal of the second line segment, Obtain the 4th judged result；

Judge the vertical coordinate of terminal of described first line segment whether more than the vertical coordinate of terminal of the second line segment, Obtain the 5th judged result；

Judge that the number of the number of the marginal point of the first line segment and the marginal point of described second line segment is the biggest In the 4th predetermined threshold value, it is thus achieved that the 6th judged result；

When described 4th judged result, described 5th judged result and described 6th judged result are when being, Using described first line segment, described second line segment combination as candidate's bright dark side boundary line.

Method the most according to claim 1, it is characterised in that described according to vehicle headlight beam Feature, determines that bright dark side boundary line includes in the candidate's bright dark side boundary line set obtained:

Obtain the response value in each candidate bright dark side boundary line, described response value is ranked up, by response value Big candidate bright dark side boundary line is as the bright dark side boundary line of vehicle head lamp；

Wherein, the response value in each candidate of described acquisition bright dark side boundary line includes:

Obtain in candidate's bright dark side boundary line the meansigma methods of the gradient modulus value of the marginal point of the line segment of first direction with And the meansigma methods of the gradient modulus value of the marginal point of the line segment of second direction；

Obtain the meansigma methods of the gradient modulus value of the marginal point of the line segment of described first direction and described second direction Line segment marginal point gradient modulus value meansigma methods and value as the response in this candidate bright dark side boundary line Value.

Method the most according to claim 5, it is characterised in that obtaining candidate's bright dark side boundary line group In conjunction before the response value in each bright dark side boundary line, described method also includes:

Utilize texture region feature and/or regional luminance feature, further screening candidate's bright dark side boundary line.

Method the most according to claim 1, it is characterised in that described method also includes:

Export positional information and/or the water in this bright dark side boundary line of composition of turning point corresponding to bright dark side boundary line Flat bright dark side boundary line and the positional information in oblique bright dark side boundary line.

8. the bright dark border line detector of vehicle head lamp based on image, it is characterised in that described Device includes:

Image acquiring unit, for obtaining the irradiation image of vehicle head lamp；

Edge detection unit, for carrying out rim detection to described irradiation image, it is thus achieved that edge image；

Candidate combinations chooses unit, and in described edge image, detection acquisition meets first party respectively To the first line segment aggregate and meet the second line segment aggregate of second direction, based on described first line segment aggregate Obtain with described second line segment aggregate and meet pre-conditioned candidate bright dark side boundary line set；Wherein, described Direction, first direction correspondence level bright dark side boundary line, the corresponding direction, oblique bright dark side boundary line of described second direction； One candidate's bright dark side boundary line is by a line segment in described first line segment aggregate and described second line segment aggregate In line segment composition；

Determine unit, for according to vehicle headlight beam feature, at the candidate's bright dark side boundary line collection obtained Conjunction determines the bright dark side boundary line of vehicle head lamp.

Device the most according to claim 8, it is characterised in that described candidate combinations chooses unit bag Unit is chosen in vinculum section detector unit and combination, and described Line segment detection unit includes the first line segment detector unit With the second Line segment detection unit, wherein, described first line segment detector unit includes:

First gradient direction angle acquiring unit, for for each marginal point in edge image, it is thus achieved that The gradient direction angle of marginal point；

First judging unit, for judging the angle that the gradient direction angle of marginal point is corresponding with first direction Difference whether less than the first predetermined threshold value, it is thus achieved that the first judged result；

First storage unit, for when described first judged result show the gradient direction angle of marginal point with When the difference of the angle that first direction is corresponding is less than the first predetermined threshold value, preserve described marginal point to the first limit Edge point set；

Clustering processing unit, for carrying out clustering processing to the marginal point obtained, it is thus achieved that corresponding line segment；

Wherein, described second Line segment detection unit includes:

Second gradient direction angle acquiring unit, for for each marginal point in edge image, it is thus achieved that The gradient direction angle of marginal point；

Second judging unit, for judging the angle that the gradient direction angle of marginal point is corresponding with second direction Difference whether less than the second predetermined threshold value, it is thus achieved that the second judged result；

Second storage unit, for when described second judged result show the gradient direction angle of marginal point with When the difference of the angle that second direction is corresponding is less than the first predetermined threshold value, preserve described marginal point to the second limit Edge point set；

Clustering processing unit, for carrying out clustering processing to the marginal point obtained, it is thus achieved that corresponding line segment.

Device the most according to claim 9, it is characterised in that described clustering processing unit includes:

Marginal point set acquiring unit, for obtaining corresponding to each marginal point in described marginal point set Polar coordinate linear equation parameter ρ, by size to parameter ρ sort, statistics obtain there is identical parameters ρ The number of marginal point, it is corresponding with this parameter ρ that acquisition has the parameter ρ corresponding to local maximum marginal point number Marginal point set；

Sequencing unit, the marginal point in the marginal point set corresponding to each parameter ρ is ranked up；

Line segment processing unit, for the marginal point after sequence, calculating the distance between the point of adjacent two edges, Determine whether adjacent two edges point belongs to same line segment according to distance.

11. devices according to claim 8, it is characterised in that described determine unit specifically for:

Obtain the response value in each candidate bright dark side boundary line, described response value is ranked up, by response value Big candidate bright dark side boundary line is as the bright dark side boundary line of vehicle head lamp；Wherein, each candidate of described acquisition The response value in bright dark side boundary line includes: obtain the marginal point of the line segment of first direction in candidate's bright dark side boundary line The meansigma methods of gradient modulus value and the meansigma methods of gradient modulus value of marginal point of line segment of second direction；Obtain Obtain meansigma methods and the line segment of described second direction of the gradient modulus value of the marginal point of the line segment of described first direction Marginal point gradient modulus value meansigma methods and value as the response value in this candidate bright dark side boundary line.

12. 1 kinds of vehicle head lamp bright dark side boundary line based on image detecting systems, it is characterised in that institute The system of stating includes in-vehicle camera and vehicle head lamp bright dark border line detector, wherein:

Described in-vehicle camera is for irradiating the irradiation image of vehicle head lamp and bright to described vehicle head lamp Dark border line detector sends the irradiation image of described vehicle head lamp；

Described vehicle head lamp bright dark border line detector is for receiving the vehicle that described in-vehicle camera sends The irradiation image of headlamp；Described irradiation image is carried out rim detection, it is thus achieved that edge image；Described In edge image, detection obtains and meets the first line segment aggregate of first direction and meet second direction respectively Second line segment aggregate, obtains meet default bar based on described first line segment aggregate and described second line segment aggregate Candidate's bright dark side boundary line set of part；Wherein, direction, described first direction correspondence level bright dark side boundary line, The corresponding direction, oblique bright dark side boundary line of described second direction；One described candidate bright dark side boundary line is by described A line segment in one line segment aggregate and a line segment composition in described second line segment aggregate；According to vehicle Headlight beam feature, determines the bright dark side of vehicle head lamp in the candidate's bright dark side boundary line set obtained Boundary line.