CN2600787Y - Device for measuring running trace of vehicle starting using digital image identification technology - Google Patents

Abstract

The utility model provides a measuring device which can measure the vehicle starting motion path with the digital image identification technology in the driver road examination. The measuring device comprise: an image collecting unit which is used to collect the tested vehicle video images; an image treating unit which is used to treat the collected video images to get the actual motion path (sliding amount) at the vehicle start-up. The utility model device which measures the vehicle starting motion path can precisely measure the sliding amount, and greatly solves the defects of the great randomicity and the incorrect results of the prior driver examination measures.

Description

The device of movement locus when utilizing the starting of digital picture recognition technology measuring vehicle

Technical field:

The utility model relates to a kind of device that utilizes digital picture recognition technology Measuring Object displacement, more particularly, the utility model relates to a kind of device of the movement locus when using the digital picture recognition technology to detect vehicle start in driver's road examination system.

Background technology:

In the car ramp starting-up process, when motor vehicle was parked on the sloping road, because the influence of gravity, vehicle can be to lower slider, and this moment, vehicle was that the hand brake by vehicle prevents what vehicle from sliding.And when vehicle start, carry out the operation of oil: igniting, engage a gear, loose clutch, step on the gas, RELEASE PARKINGBRAKE from cooperation, so if oil is when cooperation is bad, the hand brake has unclamped and vehicle does not also have enough power forward, vehicle will descend to slip, so in driver's examination, this project will be examined separately.

At present in driver's road examination process, examination person and driver be at the pilothouse of vehicle, and in the vehicle ' process, the person of getting through the examinations oneself is made Exam Evaluation to the driver to the impression of driving procedure.The drawback of this method of examination is apparent: the assessment that the examination project is made is a qualitative conclusions, rather than quantitative explanation.As the standard of estimating, make the people be easy to fairness, the fairness of examination are thrown doubt upon with oneself sensation.

The utility model content:

The purpose of this utility model is to overcome above-mentioned drawback, and a kind of device of the vehicle movement track can use the digital picture recognition technology accurately to detect vehicle start in driver's road examination system the time is provided.

Above-mentioned purpose of the present utility model is to reach by following technical scheme.

The device of the movement locus when the utility model utilizes the starting of digital picture recognition technology measuring vehicle comprises:

Image capture device is used to gather the video image of tested vehicle;

Image processing equipment is used for handled the equipment with the movement locus that obtains tested vehicle when starting by the video image that above-mentioned image capture device collected.

The device of the movement locus when the utility model utilizes the starting of digital picture recognition technology measuring vehicle, wherein said image capture device comprises:

Being fixedly installed on being used on the tested vehicle indicates the caliberating device of this vehicle shift position;

Be fixedly installed on and be used for obtaining the device of this vehicle outside the tested vehicle at the image of motion process, being arranged so that of this device can be taken in whole vehicle in its Image Acquisition visual field, and the image acquisition plane of this image acquiring device is parallel with vehicle ' ground thereon.

Described image processing equipment comprises:

Be used to be provided with the device of the inner interim coordinate system of image processing equipment;

Be used for to convert the device of digital picture to by the video image of described image capture device collection;

Be used for the data image signal that is converted to is discerned and followed the tracks of obtaining the movement locus of tested vehicle in described interim coordinate system, and and then obtain the device of the actual motion track of vehicle when starting.

The device of the movement locus when the utility model utilizes the starting of digital picture recognition technology measuring vehicle, the wherein said device that the data image signal that is converted to is discerned and followed the tracks of comprises:

Be used for being to carry out edge extracting in determined first identification in center and the tracing area with the first secondary digital picture caliberating device Position Approximate, so that obtain described caliberating device size, thus the parts of definite template size;

Be used for according to this template, in described first identification and tracing area, carry out absolute error and criterion SAD coupling according to the following equation, so that obtain the parts of the coordinate of caliberating device in the first secondary digital picture; $\mathrm{SAD}(i,j)=\underset{m=-3}{\overset{3}{\mathrm{\Σ}}}\underset{n=-3}{\overset{3}{\mathrm{\Σ}}}|M(3+m,3+n)-S(i+m,j+n)|$

Wherein, i, j are the tested point coordinates, and M is for touching the plate matrix, and S is a testing image,

Being used in the second secondary digital picture, is to carry out described SAD in second identification at center and the tracing area to mate so that obtain the parts of caliberating device coordinate in the second secondary digital picture with first caliberating device;

Be used for predicting the roughly direction of motion and the step-length of described caliberating device, so that determine the parts of next identification and tracing area according to this direction and step-length according to the coordinate of the described first secondary digital picture and the second secondary digital picture caliberating device;

Be used for repeating above-mentioned identification and following the tracks of processing, so that in once complete image processing process, write down the parts of all digital picture positions that the match is successful in corresponding identification and tracing area;

Be used for the described picture position data-switching that the match is successful is become the interior coordinate points of the inner interim coordinate system of described image processing equipment, so that obtain the parts of the movement locus of tested vehicle in this interim coordinate system.

The device of the track when the utility model utilizes the starting of digital picture recognition technology measuring vehicle, wherein said image processing equipment also comprise according to following formula, improve the parts of processing result image resolution with the sub-pix method $x_c=\frac{\underset{i,j}{\mathrm{\&Sigma;}}\frac{i}{\mathrm{SAD}(i,j)}}{\underset{i,\mathrm{<figure-callout\; id="1"\; label="j"\; filenames="Y0224661500261.png"\; state="\{\{state\}\}">j</figure-callout>}}{\mathrm{\&Sigma;}}\frac{1}{\mathrm{SAD}(i,j)}},$ $y_c=\frac{\underset{i,j}{\mathrm{\&Sigma;}}\frac{j}{\mathrm{SAD}(i,j)}}{\underset{i,\mathrm{<figure-callout\; id="1"\; label="j"\; filenames="Y0224661500261.png"\; state="\{\{state\}\}">j</figure-callout>}}{\mathrm{\&Sigma;}}\frac{1}{\mathrm{SAD}(i,j)}}$

I in the formula, j are the coordinates of all SAD less than the point of predetermined SAD one threshold value, x _cAnd y _cBe the result of inferior pixel accuracy.

The device of the movement locus when the utility model utilizes the starting of digital picture recognition technology measuring vehicle, wherein said caliberating device is to be fixedly installed on the tested vehicle, is painted with the white box that a diameter is about 20 centimetres of black cakes;

Describedly be fixedly installed on that to be used for obtaining this vehicle outside the tested vehicle be a ccd video camera at the device of the video image of motion process;

The described device that is used to be provided with the inner interim coordinate system of image processing equipment is two white crosses that are about 30 centimeter square that are fixedly installed on the rectilinear direction parallel with tested vehicle direction of motion;

The described parts that are used for the picture signal by the image capture device collection is converted to data image signal are a video frequency collection card;

The described parts that are used for the digital signal that is converted to is handled accordingly to obtain the movement locus of tested vehicle when starting to walk are a computing machine.

The device of the movement locus when the utility model utilizes the starting of digital picture recognition technology measuring vehicle, wherein said video frequency collection card is Canadian MATROX METEO II type image pick-up card;

Described computing machine uses the Intel chipset, and CPU is PIII866, in save as 512M, hard disk is 40G.

The device that the utility model provides is mainly used in the car ramp starting project in driver's road examination.The method of examination in the past is that the examiner judges the movement locus of vehicle when the starting with sensation in car, i.e. sliding amount, and what lean on is experience, and does not have quantitative data, therefore allows the people that the accuracy of examination is thrown doubt upon easily; And the intervention of human factor, cause the malpractices situation in the examination process easily; Each examiner's experience also has difference to same examinee's understanding.The device of the track when the utility model utilizes the starting of digital picture recognition technology measuring vehicle has solved the problem that present driver's wire examination method has been not suitable for the current social development preferably in line with the service principle of fair, just and open.

Description of drawings:

The purpose of this utility model, advantage and other characteristics will be more clear by the detailed description below in conjunction with accompanying drawing.These accompanying drawings are:

Fig. 1 is the general structure synoptic diagram according to the vehicle movement trajectory measurement device of the utility model one embodiment;

Fig. 2 is a width of cloth photo site captured according to the vehicle movement trajectory measurement device of the utility model one embodiment;

Fig. 3 is the synoptic diagram that is arranged on the caliberating device in the tested vehicle according to the vehicle movement trajectory measurement device of the utility model one embodiment;

Fig. 4 is the process flow diagram according to the performed Digital Image Processing process of the vehicle movement trajectory measurement device of the utility model one embodiment;

Fig. 5 is the process flow diagram that utilizes the SAD algorithm identified in the Digital Image Processing process shown in Figure 4 and follow the tracks of the caliberating device in the tested vehicle;

Fig. 6 is the distribution plan that the vehicle movement trajectory measurement device according to the utility model one embodiment utilizes the SAD algorithm to obtain;

Fig. 7 and Fig. 8 are the tracks when utilizing vehicle movement trajectory measurement device of the present utility model to measure resulting vehicle start.

Embodiment:

Referring to Fig. 1.Fig. 1 is the general structure synoptic diagram according to the vehicle movement trajectory measurement device of the utility model one embodiment.As shown in the figure, vehicle movement trajectory measurement device of the present utility model mainly comprises following ingredient: video camera 1, and tested vehicle 3, process computer 5, wherein video camera 1 is fixedly mounted on the place about about 6 meters apart from ground by a mounting bracket 2.Video camera 1 couples together by a concentric cable 4 and a process computer 5.Video camera 1 is taken the travel situations of vehicle 3 in the visual field; The video frequency collection card (not shown) is converted to computing machine 5 treatable digital pictures to the vision signal that is transmitted by video camera 1; Process computer 5 is finished the calculating of digital signal, the demonstration of testing process and the storage and the transmission of testing result.Video camera 1 will be installed in the position about 6 meters apart from ground, and concrete height will decide according to the size of taking the visual field and the focal length of camera, and principle is to take in whole car in the visual field at least.And its image pickup plane is parallel with the ground that vehicle 3 travels thereon, and along slope coordinate is parallel with vehicle heading.The distance on camera and ground will accurately measure.

Referring to Fig. 2 and Fig. 3.Fig. 2 is a width of cloth photo site captured according to the vehicle movement trajectory measurement device of the utility model one embodiment.As shown in the figure, be set with a vehicle identification tag 6 at the top of tested vehicle 3.Fig. 3 shows the synoptic diagram of distinguishing mark 6 in this example, and this vehicle identification tag 6 preferably uses the A3 blank sheet of paper, and the black cake of an about 20cm of diameter is drawn at the center.Can see also that in Fig. 2 the straight line parallel with vehicle 3 travel directions is provided with two terrestrial references 7.In this example, terrestrial reference is two white crosses, and color is selected bigger with the ground colour contrast, 30 centimeter square, and white lines are wide 10 centimetres.

Distinguishing mark 6 is fixed on vehicle 3 tops, and process computer 5 is sought this sign 6 to calculate the coordinate position of vehicle 3 in the visual field in every two field picture that video camera 1 is taken; Terrestrial reference 7 fixes on the ground, is used to demarcate coordinate system and also can eliminates owing to video camera 1 rocks the identification error that causes.The installation direction of terrestrial reference 7 is parallel with vehicle 3 direct of travels, and the distance between two terrestrial references will accurately measure.Vehicle identification tag 6 and terrestrial reference 7 will carry out image-capture by video camera 1 before identification, use when being stored in the inner so that identification of computing machine 5 with file mode.

In identifying, the particular location of vehicle 3 is to identify with the process computer 5 inner interim coordinate systems of setting up, and this coordinate system will reflect the traveling-position of vehicle 3 reality on the road surface, therefore must be converted to earth coordinates the interim coordinate system of program inside.According to the accurate distance of 7 of terrestrial references, the landmark image in the identification visual field just can calculate accurate conversion coefficient.

When Flame Image Process begins, set up interim coordinate system in computing machine 5 inside according to visual field size and video camera 1 antenna height.After vehicle 3 sails the visual field into, the vision signal that the capture card (not shown) is taken video camera 1 is converted to digital picture, computing machine 5 is sought the distinguishing mark 6 that is fixed on the car body 3 in these digital pictures, the position of this sign 6 is converted to coordinate in the interim coordinate system, just can obtains the exact position of vehicle 3 at last.

Attention needs to set identified region in image recognition processes.Identified region is to specify part zone as the effective recognition zone in whole visual field, and the reason of doing like this is in order to improve program efficiency, prevents that each Flame Image Process from all seeking distinguishing mark in whole visual field.Outside this zone, no longer discern.

After more than setting is finished, directly enter identification processing procedure behind each the startup and get final product.Wait for when initial that vehicle sails into, per second is gathered 13 two field pictures, vehicle 3 one enters the visual field, and this processing procedure promptly begins to seek distinguishing mark 6, and the location records that at every turn identifies is got off, when vehicle 3 rolls visual field or identified region away from, or after having reached maximum recognition time, finish identification this time, in interim coordinate system, draw vehicle 3 driving traces, get back to original state then, wait for next time and measuring.

If used process computer master 5 processes image in the utility model, therefore its CPU and internal memory there is special requirement.The practical following configuration of present embodiment:

CPU: requirement is the Intel chipset, to satisfy the compatible requirement of image pick-up card.And, select the high CPU of dominant frequency as far as possible for use for the consideration of operation time, that select for use in this example is the CPU of PIII866, along with development of computer, can select the CPU of higher configuration for use.

Internal memory: digital picture is stored in the process computer internal memory in this device.The data volume of every width of cloth image is: 768 * 576 * 1byte=432k, per second gather 13 width of cloth images, so the data volume of per second is: 432 * 13=5.6M.Because during whole starting-up process can not surpass 30 seconds, the minimum internal memory of therefore distributing to digital picture should be: 5.6M * 30=168M, consider to prevent overflow error, be divided into the dispensing digital picture in save as 256M; To leave the internal memory of operating system and application program some in addition for, be configured to 256 * 2=512M at present.Certainly along with the development and the measurement actual needs of technology, also can adopt more jumbo internal memory.

Hard disk: for the consideration that will store intermediate images, hard disk should be selected bigger capacity for use, and selecting capacity in this example for use is the hard disk of 40G.Certainly capacity is bigger then better.

Camera: 4.5mm C/CS bayonet socket, auto iris.

Video camera: colored CCD industrial camera.

Capture card: Canadian MATROX METEO II type image pick-up card, can certainly adopt any other image pick-up card with image acquisition and translation function.

In order can successfully to measure, device of the present utility model should be noted that following item when mounted:

The auto iris control plug of camera lens will insert the aperture control socket of ccd video camera 1.Video camera 1 is fixed on the support 2, notes rainproofly and dustproof, accomplishes firm as far as possible.Apart from about 6 meters on ground, guarantee to park an automobile 3 in the visual field.The angle requirement of laying of video camera 1 is: from monitor, the base of image (long limit) is parallel to vehicle heading (road), and vehicle 3 cruisings should enter the visual field from the image left side, sail out of the visual field from the image right side.The video output of ccd video camera 1 is connected to the video inputs of capture card.Video camera 1 need connect power supply (220V/50Hz) separately.Note the correct camera lens of installing, tighten lock screw after adjusting focal length.

Terrestrial reference 7 is set up with preventing video camera 1 shake in order to demarcate the visual field.Terrestrial reference 7 will be parallel to the track to be installed in the roadside, and two monumented points are in alignment, parallel with vehicle 3 direct of travels, notes not blocked by car 3.

Vehicle mark point 6 can be attached on covering for roofs or the headstock.Automobile 3 is preferably pure color, and color is not limit, and can not have decorative pattern to disturb but overlook.Do not place article such as books and newspapers on the panel board, behind the rear seat headrest, in order to avoid cause interference to overlooking.

Referring to Fig. 4.Fig. 4 shows the process flow diagram of the performed Digital Image Processing process of the vehicle movement trajectory measurement device of the utility model one embodiment.As shown in the figure, when vehicle 3 input identified regions, begin to carry out identification and handle (S400).Carry out initialization at S410.Judge whether to begin to detect at S430.If the result is a "Yes", then advance to S460, at this step collection vehicle image, judge whether to finish to gather at S470.If the result is a "Yes", then processing procedure advances to S480 the image that collects is discerned.Program advances to S490 and preserves the video image that collects then, sends in the S495 request then and examines auditory network.If the result in S430 is a "No", then further judge whether to transmit data at S440, be "Yes" if answer, then carry out data and transmit at S450.

Referring to Fig. 5 identifying is described.Fig. 5 is the process flow diagram that utilizes the SAD algorithm identified in the Digital Image Processing process shown in Figure 4 and follow the tracks of the monumented point 6 in the tested vehicle 3.Discern the method that adopts template matches, the image (Fig. 3) of the vehicle mark 6 that grasps when promptly basis is initial mates the similar image in the visual field, thereby finishes the tracking to vehicle 3 driving traces, the movement locus when finishing vehicle launch, the i.e. calculating of sliding amount.Can trigger identifying with external signal, also can after vehicle 3 admissions, begin identification automatically.

Identifying is from S500.The manual general position of monumented point 6 that marks out in first width of cloth image at first, be the center with this position then, do edge extracting earlier one time in 20 * 20 the zone (size in zone is determined by the speed of vehicle 3 motions), obtain the size of monumented point 6, determine the size (S510) of template.Then, in this zone, do the SAD template matches with this template, obtain the coordinate of monumented point in first width of cloth image at S520; In second width of cloth image, be the center with the coordinate of the monumented point 6 that obtains in first width of cloth image, do the SAD coupling in 20 * 20 zones, just can obtain the coordinate of monumented point 6 in second width of cloth image.Judge at S530 whether above-mentioned coupling is successful, if success, then identifying advances to S540, utilizes above-mentioned two recognition results to dope the roughly direction of motion and the step-length of monumented point 6, determine the zone of search next time with this direction and step-length again, thereby reach the purpose of tracking.At S550, the picture position that the match is successful is noted, convert the coordinate points in the interim coordinate system to.After identifying is finished, draw the movement locus in the time of just can on screen, obtaining vehicle 3 and travel in coordinate system temporarily according to resulting coordinate points.At S560, according to these data, the physical process of travelling with reference to vehicle 3 by fixing algorithm, just can obtain vehicle 3 accurate maximum coasting distance, i.e. sliding amount in whole uphill starting process.This data output, image processing process is got back to original state, waits for identifying next time.If identifying is sought the image of coupling in whole identified region, then operand is bigger, can be more time-consuming.And the above-mentioned identification processing procedure that the utility model adopts utilizes the physical correlation principle of object of which movement, as long as the initial matching success, later matching process just need not carry out in whole identified region.So just can significantly reduce calculated amount, make the efficient of identifying improve a lot.

Simply introduce the SAD matching algorithm below in conjunction with Fig. 6.

According to the above description, what obtained vehicle mark point 6 touches the plate size, slides in the identification region of search with this template, adopts absolute error and criterion (SAD) as matching criteria.The SAD formula is as follows: $\mathrm{SAD}(i,j)=\underset{m=-3}{\overset{3}{\mathrm{\&Sigma;}}}\underset{n=-3}{\overset{3}{\mathrm{\&Sigma;}}}|M(3+m,3+n)-S(i+m,j+n)|$

I, j are the tested point coordinates in the formula, the center in zone to be measured just, and M is a pattern matrix, S is a testing image.This formula has been described the similarity degree of zone to be measured (7 * 7) and template, and SAD is more little, illustrates that zone to be measured is similar more to template, and (i j) might be the centre coordinate of monumented point more to have the point of more little SAD so.

Fig. 6 is above-mentioned distribution situation to the sad value in two monumented point 6 identifyings in first sub-picture and second sub-picture, and as can be seen, near monumented point 6, SAD has significantly downward spike, and this position is exactly the position of monumented point 6.

Should be appreciated that both available hardware realizations of above-mentioned image processing process of the present utility model, also available software realizes.In the present embodiment, realize with software.Relevant program is developed under the VC++6.0 the integration environment, has utilized Intel IPL2.5, OpenCV2.0 and MilLit storehouse.Use application program on the engineering based on dialog box.Operating system can be selected Windows NT 4/Windows2000 for use, to guarantee reliability of operation.Enter key frame when program is initial, can select for use in the menu this moment option finish the work the extracting described in the process, demarcation, setting sampling time, set function such as sample area, after above-mentioned work was finished, program can restart, and waited beginning to be identified.

Because the influence of external environment condition, the processing of the readability of template, edge situation, the adjusting of camera focal length, the shake of video camera and the influence of the speed of a motor vehicle can bring stochastic error to system; And the measuring error of camera height in the calibration process, measuring error, position that video camera is installed and the attitude error of subject distance and the site error that terrestrial reference is put, all can introduce fixed error to measurement result.Therefore must use proper method to improve the precision of identification.

At first to control the introducing of fixed error, when video camera 1 is installed, want the usage level instrument with the camera lens normal of assurance camera perpendicular to ground, and the edge of visual field will be parallel to the direction that vehicle 3 is advanced; When fixing terrestrial reference 7, should guarantee as far as possible that the line of centres of terrestrial reference 7 is parallel to the direction that vehicle 3 is advanced; When measuring video camera 1 height, use pedal line, what guarantee to measure is the vertical range of lens plane to ground; Equally measure terrestrial reference 7 apart from the time also to consider this point.By accurate installation and accurately measurement, can be controlled at a very little scope to fixed error.

The supression of stochastic error must realize by relevant algorithm.What adopt in this method is inferior pixel algorithm, identification position control within a pixel.It is generally acknowledged that the resolution of digital image processing system is a pixel.In fact, because the continuity of image can realize a precision below the pixel by mathematical method.Inferior pixel is exactly wherein a kind of method.Referring to Fig. 2, if we set a SAD threshold value, as 1000, think that the point less than this threshold value all may be the position of monumented point and since SAD more little might be the position of monumented point more, so the inverse of SAD as weights, adopt weighted-average method to improve resolution, formula is as follows: $x_c=\frac{\underset{i,j}{\mathrm{\&Sigma;}}\frac{i}{\mathrm{SAD}(i,j)}}{\underset{i,\mathrm{<figure-callout\; id="1"\; label="j"\; filenames="Y0224661500261.png"\; state="\{\{state\}\}">j</figure-callout>}}{\mathrm{\&Sigma;}}\frac{1}{\mathrm{SAD}(i,j)}},$ $y_c=\frac{\underset{i,j}{\mathrm{\&Sigma;}}\frac{j}{\mathrm{SAD}(i,j)}}{\underset{i,\mathrm{<figure-callout\; id="1"\; label="j"\; filenames="Y0224661500261.png"\; state="\{\{state\}\}">j</figure-callout>}}{\mathrm{\&Sigma;}}\frac{1}{\mathrm{SAD}(i,j)}}$

I in the formula, j are the coordinates of all SAD less than the point of threshold value, x _cAnd y _cBe the result of inferior pixel accuracy.Can extrapolate a pixel according to the size of image and be about as much as about the 8mm of actual coordinate, the precision that is to say identification is in 1cm.

Fig. 7 and Fig. 8 show according to device of the present utility model resulting measurement result in the simulating vehicle reversing process.The concrete driving process of vehicle is: admission-parking-measurement initial position-reversing-parking-measurement end position-to overtake, and can know from the curve tendency and see this process.The reversing amount that measures shown in Fig. 7 is 278mm, and is 274mm from the end value that identification obtains.4 millimeters of error have been verified the precision that algorithm proposes.Measuring process shown in Fig. 8 is with top the same, and the result who measures is 412mm, and the result that identification obtains is 406mm, error 6mm.

In the actual experiment process, analyzing lot of data can obtain, the identification error basic controlling measuring distance 2% in, so satisfy the requirement of actual measurement fully.

Device of the present utility model, consider the singularity in the actual examination process of vehicle drive, adopted the digital picture recognition methods, thereby finished carrying out automatically of whole examination process, do not need installation of additional equipment (only need obtain laying highly of monumented point 6 gets final product) onboard, and, make measuring accuracy reach very high degree, the requirement of complete realistic examination by perfect algorithm design.

Device of the present utility model has not only been finished experimentation, and has carried out on-the-spot demonstration in relevant examination field, has obtained an examination expert's approval, can put in the middle of the actual use fully.

Though with accompanying drawing the utility model is had been described in detail in conjunction with the embodiments above; but be to be understood that; concerning the those skilled in the art of this area; protection domain of the present utility model is not limited thereto; under the situation that does not deviate from spirit and scope of the present utility model; can also make many variations and modification to the utility model, all such modifications and variation all should be considered to fall into by in the appended claims institute restricted portion.

Claims (6)

1, a kind of device that utilizes the movement locus in digital picture recognition technology measuring vehicle when starting, it is characterized in that: this device comprises:

Be used to gather the image capture device of the video image of tested vehicle;

Be used for the video image that is collected by above-mentioned image capture device is handled, so that obtain the image processing equipment of the movement locus in tested vehicle when starting.

The device of the movement locus when 2, measuring vehicle as claimed in claim 1 starts to walk, it is characterized in that: described image capture device comprises:

Being fixedly installed on being used on the tested vehicle indicates this vehicle movement position calibration device;

Be fixedly installed on and be used for obtaining the device of this vehicle outside the tested vehicle at the video image of motion process, this device is configured such that can take in whole vehicle in its Image Acquisition visual field, and the image acquisition plane of this image acquiring device is parallel with vehicle ' ground thereon;

Described image processing equipment comprises: the device that is used to be provided with the inner interim coordinate system of image processing equipment;

Be used for to convert the device of digital picture to by the video image of described image capture device collection;

Be used for the data image signal that is converted to is discerned and followed the tracks of obtaining the movement locus of tested vehicle at described interim coordinate system, and and then obtain the device of the actual motion track of vehicle when starting.

The device of the movement locus when 3, measuring vehicle as claimed in claim 2 starts to walk, it is characterized in that: the described device that the data image signal that is converted to is discerned and followed the tracks of comprises:

Be used in the Position Approximate with the first secondary digital picture caliberating device is determined first identification in center and tracing area, carrying out edge extracting, so that obtain described caliberating device size, thus the parts of definite template size;

Be used for according to this template, in described first identification and tracing area, carry out absolute error and criterion SAD coupling according to the following equation, so that obtain the parts of the coordinate of caliberating device in the first secondary digital picture; $\mathrm{SAD}(i,j)=\underset{m=-3}{\overset{3}{\mathrm{\&Sigma;}}}\underset{n=-3}{\overset{3}{\mathrm{\&Sigma;}}}|M(3+m,3+n)-S(i+m,j+n)|$

Wherein, i, j are the tested point coordinates, and M is for touching the plate matrix, and S is a testing image,

Be used in the second secondary digital picture with first caliberating device is second identification at center and tracing area, carrying out described SAD and mate, so that obtain the parts of caliberating device coordinate in the second secondary digital picture;

Be used for predicting the roughly direction of motion and the step-length of described caliberating device, so that determine the parts of next identification and tracing area according to this direction and step-length according to the coordinate of the described first secondary digital picture and the second secondary digital picture caliberating device;

Be used for repeating above-mentioned identification and following the tracks of processing, so that in once complete image processing process, write down the parts of all digital picture positions that the match is successful in corresponding identification and tracing area;

Be used for the described picture position data-switching that the match is successful is become the interior coordinate points of the inner interim coordinate system of described image processing equipment, so that obtain the parts of the movement locus of tested vehicle in this interim coordinate system.

4, as the device of the track during each described measuring vehicle starting among the claim 1-3, it is characterized in that: described image processing equipment also comprises according to following formula, improves the parts of processing result image resolution with the sub-pix method $x_c=\frac{\underset{i,j}{\mathrm{\&Sigma;}}\frac{i}{\mathrm{SAD}(i,j)}}{\underset{i,j}{\mathrm{\&Sigma;}}\frac{1}{\mathrm{SAD}(i,j)}},$ $y_c=\frac{\underset{i,j}{\mathrm{\&Sigma;}}\frac{j}{\mathrm{SAD}(i,j)}}{\underset{i,j}{\mathrm{\&Sigma;}}\frac{1}{\mathrm{SAD}(i,j)}}$

I in the formula, j are the coordinates of all SAD less than the point of a predetermined SAD threshold value, x _cAnd y _cBe the result of inferior pixel accuracy.

5, as the device of the movement locus during each described measuring vehicle starting among the claim 1-3, it is characterized in that: described caliberating device is to be fixedly installed on the tested vehicle, is painted with the white box that a diameter is about 20 centimetres of black cakes;

Describedly be fixedly installed on that to be used for obtaining this vehicle outside the tested vehicle be a ccd video camera at the device of the video image of motion process;

The described device that is used to be provided with the inner interim coordinate system of image processing equipment is two white crosses that are about 30 centimeter square that are fixedly installed on the straight line parallel with tested vehicle direction of motion;

The described parts that are used for the picture signal by the image capture device collection is converted to data image signal are a video frequency collection card;

It is described that to be used for handling with the parts that obtain the tested vehicle movement locus accordingly to the digital signal that is converted to be a computing machine.

The device of the movement locus when 6, measuring vehicle as claimed in claim 5 starts to walk, it is characterized in that: described video frequency collection card is Canadian MATROX METEO II type image pick-up card; Described computing machine uses the Intel chipset, and CPU is PIII866, in save as 512M, hard disk is 40G.

Images