CN105070098B - A kind of vehicle distance detecting method based on car plate position - Google Patents

Abstract

The present invention provides a kind of vehicle distance detecting method based on car plate position, including obtains car plate detection grader and target location depthmeter；Distance of the car plate position away from image lower boundary in car plate detection grader and image based on acquisition, positions objects ahead vehicle；According to the car plate position of target vehicle in previous frame image, the car plate position of target vehicle in current frame image is predicted；Obtain distance of the target vehicle away from image lower boundary in current frame image, the target location depthmeter based on acquisition, acquisition actual range of the target vehicle away from vehicle-mounted vidicon of tabling look-up；Actual range of the target vehicle away from vehicle-mounted vidicon is subtracted into distance of the vehicle-mounted vidicon away from this car headstock, distance of the target vehicle away from this car headstock is obtained.The present invention uses machine vision learning algorithm, is accurately positioned front truck car plate position, and the resetting of car plate is carried out based on target following technology, and the demarcation of target location depthmeter is carried out based on linear interpolation algorithm；Faster, spacing calculates more accurate algorithm speed.

Description

A kind of vehicle distance detecting method based on car plate position

Technical field

The present invention relates to safe driving technical field, specifically a kind of vehicle distance detecting method based on car plate position.

Background technology

In daily motor vehicle driving, enough spacings are the most effective modes for avoiding rear-end collision with front truck holding.And For the judgement of spacing, mainly estimated and obtained by the experience of driver, this mode has wretched insufficiency：First, drive The sitting posture of member and the difference at visual angle, the result of range estimation can have relatively large deviation, especially on a highway, due to speed mistake It hurry up, the accurate spacing of acquisition can not be estimated at all；Secondly, drive often to make driver attention not concentrate for a long time, easily The spacing or spacing ignored with front truck judge there is larger error, and then trigger traffic accident.

In recent years, some spacing detection techniques are occurred in that, mainly there are following a few classes：

(1) physical distance measurement technology is based on, such technology mainly by launching and receiving ultrasonic wave or infrared laser line, is obtained Obtain the distance with preceding vehicle.There is more deficiency in this technology：Equipment cost is high, and telemeasurement error is larger, Yi Shouqian Square barrier influences and causes flase drop, and multiple vehicles can exist interfering with each other when using simultaneously.

(2) video processing technique is based on, such as Chinese patent application CN104392629A discloses a kind of side for detecting spacing Method and device, Chinese patent CN101941438B disclose a kind of intelligent measuring and controlling device of safe distance between vehicles and method, such technology master The feature related to front vehicles is obtained by video image processing technology, on each two field picture of video, be reflected according to the depth of field Firing table or three-dimensional measurement technology, obtain the spacing with front truck.The advantage of this method is low cost, actively measurement, adaptability Extensively, it has the disadvantage that algorithm is more complicated, and positioning front vehicles are inaccurate, and spacing error calculated is larger.

The content of the invention

It is an object of the invention to provide a kind of algorithm speed faster, spacing calculate the more accurate car based on car plate position Distance detecting method.

The technical scheme is that：

A kind of vehicle distance detecting method based on car plate position, comprises the following steps：

(1) car plate detection grader and target location depthmeter are obtained, the target location depthmeter is used to record target In the image that actual range and vehicle-mounted vidicon away from vehicle-mounted vidicon are gathered between distance of the target location away from image lower boundary Corresponding relation；

(2) judge whether to need to detect positioning objects ahead vehicle again, if so, step (3) is then performed, if it is not, then holding Row step (4)；

(3) distance of the car plate position away from image lower boundary, positioning in car plate detection grader and image based on acquisition Objects ahead vehicle；

(4) car plate of objects ahead vehicle is tracked, i.e., according to the car plate position of target vehicle in previous frame image, prediction is worked as The car plate position of target vehicle in prior image frame；

(5) distance of the car plate position away from image lower boundary of target vehicle in current frame image, the mesh based on acquisition are obtained Cursor position depthmeter, acquisition actual range of the target vehicle away from vehicle-mounted vidicon of tabling look-up；

(6) by actual range of the target vehicle away from vehicle-mounted vidicon subtract vehicle-mounted vidicon away from this car headstock away from From obtaining distance of the target vehicle away from this car headstock.

In the described vehicle distance detecting method based on car plate position, step (1), the acquisition car plate detection grader, bag Include：

The license plate image of different distance is used as training positive sample in a, positive and negative 30 degree of collection front；

B, based on hog features and adaboost learning algorithms training car plate detection grader file.

In the described vehicle distance detecting method based on car plate position, step (1), acquisition target location depthmeter, bag Include：

A, in right ahead, place a white rectangle cardboard respectively every a segment distance, use vehicle-mounted pick-up Machine gathers the image of all rectangle cardboards；

Each distance of the rectangular paper Board position away from image lower boundary in b, calculating image；

C, by each rectangle cardboard position in actual range and image of each rectangle cardboard away from vehicle-mounted vidicon Put the distance away from image lower boundary and substitute into below equation, obtain corresponding proportionate relationship coefficient：

Wherein, Δ Z_iRepresent i-th of actual range of the rectangle cardboard away from vehicle-mounted vidicon, Δ D_iRepresent in image i-th Distance of the rectangular paper Board position away from image lower boundary, W_iRepresent corresponding proportionate relationship coefficient；

D, using following linear interpolation formula, obtain distance correspondence of each target location away from image lower boundary in image Proportionate relationship coefficient：

Wherein, W_jRepresent image in j-th of target location away from image lower boundary apart from D_jCorresponding proportionate relationship coefficient, W_i0Represent image in j-th of target location of distance it is nearest above that distance of rectangular paper Board position away from image lower boundary D_i0Corresponding proportionate relationship coefficient, W_i1Represent in image j-th of target location of distance it is nearest below that rectangle cardboard position Put away from image lower boundary apart from D_i1Corresponding proportionate relationship coefficient；

E, according to Z_j=W_j*D_j, the target location of each in image is calculated away from image lower boundary apart from D_jCorresponding mesh The actual range Z of gauge length vehicle-mounted vidicon_j, corresponding relation is stored in the depthmeter of target location.

The described vehicle distance detecting method based on car plate position, step (3), including：

A, the car plate detection grader based on acquisition, detection front track whether there is car plate, i.e. vehicle；

B, judge obtain vehicle whether be located at current lane, if so, being considered as effective vehicle, if it is not, being considered as unproductive vehicle；

If c, an effective vehicle is only existed, as target vehicle；If there is multiple effective vehicles, then select Car plate position in image is selected away from the minimum effective vehicle of image lower boundary distance as final target vehicle.

In the described vehicle distance detecting method based on car plate position, step (4), target carriage in the image according to previous frame Car plate position, prediction current frame image in target vehicle car plate position, including：

A, in previous frame image, the car plate central point of target vehicle is saved as into target point, and by the car of target vehicle The board band of position saves as the plate template of target vehicle in current frame image after being extended to surrounding；

B, target point is tracked, based on kalman Filter Principles, predicts that the car plate central point of target vehicle is being worked as Position in prior image frame；

Mesh in c, position and previous frame image of the car plate central point in current frame image of target vehicle based on prediction The car plate size of vehicle is marked, according to below equation, the Plate searching rectangular area of target vehicle in current frame image is obtained：

Wherein, rect.x, rect.y represent the Plate searching rectangular area rect of target vehicle in current frame image respectively Upper left corner abscissa and ordinate, center.x, center.y represent that the car plate central point of target vehicle of prediction exists respectively Position abscissa and ordinate in current frame image, rect.width, rect.height represent mesh in current frame image respectively The Plate searching rectangular area rect of vehicle width and height is marked, car_wieth, car_height represent previous frame figure respectively The car plate width and height of target vehicle as in；

D, in current frame image in the Plate searching rectangular area of target vehicle, scanned for time using plate template Go through, to each traversal position, according to below equation, calculate the confidence level that the traversal position belongs to target vehicle car plate position, The corresponding traversal position of maximum confidence is selected as the car plate position of target vehicle in current frame image：

Wherein, conf_ijRepresent that current traversal position belongs to the confidence level of target vehicle car plate position, N represents present frame figure The pixel quantity of the plate template of target vehicle as in, M (x, y) represents that the plate template of target vehicle in current frame image exists The grey scale pixel value at (x, y) place, f (i+x, j+y) represented in current frame image in the Plate searching rectangular area of target vehicle, On the basis of currently traveling through position top left co-ordinate (i, j) place, to offset the grey scale pixel value at (x, y) place.

In the described vehicle distance detecting method based on car plate position, step (5), target carriage in the acquisition current frame image Distance of the car plate position away from image lower boundary, including：

A, according to below equation, obtain the binary image bin of the car plate band of position of target vehicle in current frame image (x, y)：

Wherein, f (x, y) represents the grey scale pixel value at the car plate band of position (x, y) place of target vehicle in current frame image, f(x_i, y_i) grey scale pixel value in N neighborhoods centered on (x, y) is represented, n represents the picture in the N neighborhoods centered on (x, y) Plain number, T represents binary-state threshold；

B, the binary image to acquisition carry out morphology operations, remove interference；

C, the car plate center position according to target vehicle in below equation acquisition current frame image：

Wherein, center.x, center.y represent the car plate center position of target vehicle in current frame image respectively Abscissa and ordinate, bin (xi, yi) represent (x in binary image_i, y_i) place grey scale pixel value；

D, distance of the car plate center position away from image lower boundary for calculating target vehicle in current frame image.

The described vehicle distance detecting method based on car plate position, also includes：When the target vehicle away from this car headstock away from During from less than default safe distance, audio alert is carried out.

As shown from the above technical solution, the present invention uses machine vision learning algorithm, is accurately positioned front truck car plate position, base The resetting of car plate is carried out in target following technology, the demarcation of target location depthmeter is carried out based on linear interpolation algorithm；With Existing method is compared, and faster, spacing calculates more accurate algorithm speed.

Brief description of the drawings

Fig. 1 is flow chart of the method for the present invention；

Fig. 2 is the training positive sample image that car plate detects grader；

Fig. 3 is car plate Detection results figure；

Fig. 4 is the binary image of the car plate band of position；

Fig. 5 is car plate center position figure.

Embodiment

Below in conjunction with the accompanying drawings the present invention is further illustrated with specific embodiment.

As shown in figure 1, a kind of vehicle distance detecting method based on car plate position, comprises the following steps：

S1, judge whether to need to initialize systematic parameter, if it is desired, enter step S2, otherwise, into step S3.

S2, initialization systematic parameter, systematic parameter include car plate and detect grader and target location depthmeter, specific demarcation Step is as follows：

S21, acquisition car plate detection grader, are comprised the following steps that：

The license plate image of different distance is as training positive sample in S211, positive and negative 30 degree of collection vehicle front, such as Fig. 2 institutes Show；

S212, based on hog features and adaboost learning algorithms training car plate grader file；Relevant hog features are determined Justice may refer to document：Histograms of Oriented Gradients for Human Detection, Navneet Dalal, Bill Triggs, CVPR2005.

S22, acquisition target location depthmeter：

It can be seen from the image-forming principle of video camera, when object distance video camera is more remote, its position of imaging in the picture is just Can range image lower boundary it is more remote, simultaneously as uncertain error when equipment production error in itself and IMAQ, can make Obtain this corresponding relation and become a kind of Nonlinear Mapping relation, conventional method can not obtain accurate calibration value, therefore, Wo Mengen According to " black box " are theoretical, and this Nonlinear Mapping relation is expressed as following formula：

Δ Z=W* Δs D

Wherein, for unknown, " black box " variables, Δ Z is actual range of the object away from video camera to W, and Δ D is in image Pixel distance of the object space away from image lower boundary, the i.e. offset of objects in images position.

Now, the process of demarcation becomes the target location in actual range Δ Z of the known target away from video camera and image Offset Δ D in the case of, obtain corresponding proportionate relationship coefficient W, comprise the following steps that：

S221, in right ahead, place a white long narrow rectangular paperboard respectively every 5 meters, use vehicle-mounted pick-up Machine gathers the image of all rectangle cardboards.

S222, each distance of the rectangular paper Board position away from image lower boundary in image is calculated, be each long in image The offset of rectangular paperboard position.

S223, each rectangle cardboard in actual range and image of each rectangle cardboard away from vehicle-mounted vidicon The offset of position substitutes into Δ Z_i=W_i*ΔD_i, obtain corresponding " black box " variate-values W_i, wherein i=1,2,3 ..., n.

S224, offset and " pair between black box " variables according to target location in target actual range, image It should be related to, by the theoretical formula of following linear interpolation, obtain the corresponding proportionate relationship of offset of each target location in image Coefficient：

Wherein, W_iFor the corresponding proportionate relationship coefficient of offset of current goal position in image, W_i0、W_i1For image middle-range The corresponding proportionate relationship coefficient of offset of the up and down two rectangular paper Board positions nearest from current goal position, D_jFor image The offset of middle current goal position, D_i0For in image apart from current goal position it is nearest above that rectangular paper Board position Offset, D is the difference of nearest apart from the current goal position offset of two rectangular paper Board positions up and down in image.

S225, according to Z_j=W_j*D_j, the corresponding actual range of offset of the position of each in image is calculated, by correspondence Relation is stored in the depthmeter of target location.

S3, judge whether to need to detect front vehicles, if it is desired, enter step S4, otherwise, into step S5.

S4, detection front vehicles, judge that the front of current lane whether there is target vehicle, effect as shown in figure 3, tool Body step is as follows：

S41, the car plate detection grader obtained based on step S2, detection front track whether there is car plate, i.e. vehicle；

S42, judge obtain vehicle whether be located at current lane, if not, being considered as unproductive vehicle, if be considered as Effective vehicle；

If S43, an effective vehicle is only existed, as target vehicle；If there is multiple effective vehicles, choosing The position of its car plate in the picture is selected away from the minimum effective vehicle of image lower boundary distance as final target vehicle.

S5, tracking objects ahead vehicle, the car plate position in previous frame image, prediction car plate is in current frame image Position, it is to avoid the detection positioning of car plate that each two field picture is all repeated, comprise the following steps that：

S51, renewal tracking data, mainly in previous frame image, preserve car plate central point as tracking target point, and The plate template of current frame image is saved as after the car plate band of position is extended to surrounding；

S52, target point is tracked, is based primarily upon kalman Filter Principles, prediction car plate central point is in present frame Position in image；

S53, selection Plate searching rectangular area, are mainly based upon in the car plate center position and previous frame image of prediction Car plate size, according to below equation, obtain the approximate region scope rect of car plate in current frame image：

Wherein, rect.x, rect.y represent that the upper left corner of Plate searching rectangular area rect in current frame image is horizontal respectively Coordinate and ordinate, center.x, center.y represent that position of the car plate central point of prediction in current frame image is horizontal respectively Coordinate and ordinate, rect.width, rect.height represent rect width and height, car_width, car_ respectively Height represents the width and height of the car plate obtained in previous frame image respectively.

S54, car plate position is accurately positioned, main method is：In the Plate searching rectangular area of current frame image, use Plate template scans for traversal, each traversal position, according to below equation, calculates the traversal position and belongs to car plate position Confidence level, selects the corresponding traversal position of maximum confidence as the optimal location of car plate in current frame image：

Wherein, conf_ijRepresent that current traversal position belongs to the confidence level of car plate position, N represents the car plate of current frame image The pixel quantity of template, M (x, y) represents gray value of the plate template at pixel (x, y) place, and f (i+x, j+y) is represented in present frame In the Plate searching rectangular area of image, with the basis of currently traveling through position top left co-ordinate (i, j) place, skew (x, y) place Grey scale pixel value.

Car plate center position is away under image in S6, the offset for obtaining target vehicle, mainly acquisition current frame image The distance on border, is comprised the following steps that：

S61, according to below equation, obtain the binary image bin (x, y) of the car plate band of position in current frame image, effect Fruit is as shown in Figure 4：

Wherein, f (x, y) is the grey scale pixel value at the car plate band of position (x, y) place in current frame image, f (x_i, y_i) be with The grey scale pixel value in N neighborhoods centered on (x, y), n is the number of pixels in the N neighborhoods centered on (x, y), and T is two-value Change threshold value.

S62, morphology operations, remove the less interference of area.

S63, according to below equation, obtain the car plate center position center in current frame image, effect such as Fig. 5 institutes Show：

Wherein, center.x, center.y are the abscissa and vertical seat of car plate center position in current frame image respectively Mark, bin (x_i, y_i) it is (x in binary image_i, y_i) place grey scale pixel value.

Distance of the car plate center position away from image lower boundary in S64, calculating current frame image, as in current frame image The offset of target vehicle position.

S7, the actual range for obtaining target vehicle, are comprised the following steps that：

S71, the offset according to the step S64 target vehicle positions obtained, the target location obtained according to step S22 are deep Table is spent, the corresponding depth of acquisition of tabling look-up, is distance of the objects ahead vehicle away from vehicle-mounted vidicon.

S72, correction target vehicle distance, because vehicle-mounted vidicon is typically all to be placed on vehicle interior, what step S71 was obtained Target vehicle distance contains distance of this car headstock away from vehicle-mounted vidicon, therefore, according to below equation, obtains this final car Headstock is away from the front truck tailstock apart from Z：

Z=Z_all-Z_car

Wherein, Z_allIt is distance of the vehicle-mounted vidicon away from front truck, Z_carIt is distance of the vehicle-mounted vidicon away from this car headstock.

S8, audio alert, according to the safe distance set in advance, if actual spacing is less than safe distance, voice reporting It is alert, remind driver to adjust spacing, safe driving in time.

The above embodiment is only that the preferred embodiment of the present invention is described, not to the model of the present invention Enclose and be defined, on the premise of design spirit of the present invention is not departed from, technical side of the those of ordinary skill in the art to the present invention In various modifications and improvement that case is made, the protection domain that claims of the present invention determination all should be fallen into.

Claims (6)

1. a kind of vehicle distance detecting method based on car plate position, it is characterised in that comprise the following steps：

(1) car plate detection grader and target location depthmeter are obtained, the target location depthmeter is used to record target away from car Carry pair between distance of the target location away from image lower boundary in the actual range of video camera and the image of vehicle-mounted vidicon collection It should be related to；

(2) judge whether to need to detect positioning objects ahead vehicle again, if so, step (3) is then performed, if it is not, then performing step Suddenly (4)；

(3) distance of the car plate position away from image lower boundary, positioning front in car plate detection grader and image based on acquisition Target vehicle；

(4) car plate of objects ahead vehicle is tracked, i.e., according to the car plate position of target vehicle in previous frame image, predicts present frame The car plate position of target vehicle in image；

(5) distance of the car plate position away from image lower boundary of target vehicle in current frame image is obtained, the target position based on acquisition Depthmeter is put, acquisition actual range of the target vehicle away from vehicle-mounted vidicon of tabling look-up；

(6) actual range of the target vehicle away from vehicle-mounted vidicon is subtracted into distance of the vehicle-mounted vidicon away from this car headstock, i.e., Obtain distance of the target vehicle away from this car headstock；

In step (1), acquisition target location depthmeter, including：

A, in right ahead, place a white rectangle cardboard respectively every a segment distance, adopted using vehicle-mounted vidicon Collect the image of all rectangle cardboards；

Each distance of the rectangular paper Board position away from image lower boundary in b, calculating image；

C, by each rectangular paper Board position in actual range and image of each rectangle cardboard away from vehicle-mounted vidicon away from The distance of image lower boundary substitutes into below equation, obtains corresponding proportionate relationship coefficient：

$W_i=\frac{{\mathrm{\ΔZ}}_i}{{\mathrm{\ΔD}}_i}$

Wherein, Δ Z_iRepresent i-th of actual range of the rectangle cardboard away from vehicle-mounted vidicon, Δ D_iRepresent image in i-th it is rectangular Distance of the shape cardboard position away from image lower boundary, W_iRepresent corresponding proportionate relationship coefficient；

D, using following linear interpolation formula, obtain the corresponding ratio of distance of each target location away from image lower boundary in image Example coefficient of relationship：

$\left\{\begin{array}{c}{W}_j=W_{i0}*s+W_{i1}*(1-s)\\ s=\frac{D_j-D_{i0}}{D_{i0}-D_{i1}}\end{array}\right.$

Wherein, W_jRepresent image in j-th of target location away from image lower boundary apart from D_jCorresponding proportionate relationship coefficient, W_i0Table In diagram picture j-th of target location of distance it is nearest above that rectangular paper Board position away from image lower boundary apart from D_i0Correspondence Proportionate relationship coefficient, W_i1Represent image in j-th of target location of distance it is nearest below that rectangular paper Board position away from figure As lower boundary is apart from D_i1Corresponding proportionate relationship coefficient；

E, according to Z_j=W_j*D_j, the target location of each in image is calculated away from image lower boundary apart from D_jCorresponding target away from The actual range Z of vehicle-mounted vidicon_j, corresponding relation is stored in the depthmeter of target location.

2. the vehicle distance detecting method according to claim 1 based on car plate position, it is characterised in that described in step (1) Car plate detection grader is obtained, including：

The license plate image of different distance is used as training positive sample in a, positive and negative 30 degree of collection front；

B, based on hog features and adaboost learning algorithms training car plate detection grader file.

3. the vehicle distance detecting method according to claim 1 based on car plate position, it is characterised in that step (3), including：

A, the car plate detection grader based on acquisition, detection front track whether there is car plate, i.e. vehicle；

B, judge obtain vehicle whether be located at current lane, if so, being considered as effective vehicle, if it is not, being considered as unproductive vehicle；

If c, an effective vehicle is only existed, as target vehicle；If there is multiple effective vehicles, then selection is schemed Car plate position is used as final target vehicle away from the minimum effective vehicle of image lower boundary distance as in.

4. the vehicle distance detecting method according to claim 1 based on car plate position, it is characterised in that described in step (4) According to the car plate position of target vehicle in previous frame image, the car plate position of target vehicle in current frame image is predicted, including：

A, in previous frame image, the car plate central point of target vehicle is saved as into target point, and by the car plate position of target vehicle Put the plate template that target vehicle in current frame image is saved as after region extends to surrounding；

B, target point is tracked, based on kalman Filter Principles, predicts the car plate central point of target vehicle in present frame Position in image；

Target carriage in c, position and previous frame image of the car plate central point in current frame image of target vehicle based on prediction Car plate size, according to below equation, obtain the Plate searching rectangular area of target vehicle in current frame image：

$\left\{\begin{array}{c}rect.x=center.x-car\_width*0.7\\ rect.y=center.y-car\_height\\ rect.width=car\_width*1.4\\ rect.height=car\_height*2\end{array}\right.$

Wherein, rect.x, rect.y represent the Plate searching rectangular area rect of target vehicle in a current frame image left side respectively Upper angle abscissa and ordinate, center.x, center.y represent the car plate central point of the target vehicle of prediction current respectively Position abscissa and ordinate in two field picture, rect.width, rect.height represent target carriage in current frame image respectively Plate searching rectangular area rect width and height, car_width, car_height are represented in previous frame image respectively The car plate width and height of target vehicle；

D, in current frame image in the Plate searching rectangular area of target vehicle, scan for traversal using plate template, it is right Each traversal position, according to below equation, calculates the confidence level that the traversal position belongs to target vehicle car plate position, selection is most The big corresponding traversal position of confidence level as target vehicle in current frame image car plate position：

${\mathrm{conf}}_{ij}=\frac{1}{255*N}\Σ\Σf(i+x,j+y)-M(x,y)$

Wherein, conf_ijRepresent that current traversal position belongs to the confidence level of target vehicle car plate position, N is represented in current frame image The pixel quantity of the plate template of target vehicle, M (x, y) represents the plate template of target vehicle in current frame image at (x, y) The grey scale pixel value at place, f (i+x, j+y) is represented in current frame image in the Plate searching rectangular area of target vehicle, to work as On the basis of preceding traversal position top left co-ordinate (i, j) place, the grey scale pixel value at (x, y) place is offset.

5. the vehicle distance detecting method according to claim 1 based on car plate position, it is characterised in that described in step (5) Distance of the car plate position away from image lower boundary of target vehicle in current frame image is obtained, including：

A, according to below equation, obtain the car plate band of position of target vehicle in current frame image binary image bin (x, y)：

$bin(x,y)=\left\{\begin{array}{cc}255& f(x,y)-\frac{1}{n}{\Σ}_{i=1}^{N}f(x_i,y_i)\≥T\\ 0& f(x,y)-\frac{1}{n}{\Σ}_{i=1}^{N}f(x_i,y_i)\≥T\end{array}\right.$

Wherein, f (x, y) represents the grey scale pixel value at the car plate band of position (x, y) place of target vehicle in current frame image, f (x_i, y_i) grey scale pixel value in N neighborhoods centered on (x, y) is represented, n represents the pixel in the N neighborhoods centered on (x, y) Number, T represents binary-state threshold；

B, the binary image to acquisition carry out morphology operations, remove interference；

C, the car plate center position according to target vehicle in below equation acquisition current frame image：

$\left\{\begin{array}{c}center.x=\frac{\Σx_i\·*bin(x_i,y_i)}{\Σbin(x_i,y_i)}\\ center.y=\frac{\Σy_i*bin(x_i,y_i)}{\Σbin(x_i,y_i)}\end{array}\right.$

Wherein, center.x, center.y represent the horizontal seat of the car plate center position of target vehicle in current frame image respectively Mark and ordinate, bin (x_i, y_i) represent (x in binary image_i, y_i) place grey scale pixel value；

D, distance of the car plate center position away from image lower boundary for calculating target vehicle in current frame image.

6. the vehicle distance detecting method according to claim 1 based on car plate position, it is characterised in that also include：When described When distance of the target vehicle away from this car headstock is less than default safe distance, audio alert is carried out.