CN101739827B - Vehicle detecting and tracking method and device - Google Patents
Vehicle detecting and tracking method and device Download PDFInfo
- Publication number
- CN101739827B CN101739827B CN200910238267.3A CN200910238267A CN101739827B CN 101739827 B CN101739827 B CN 101739827B CN 200910238267 A CN200910238267 A CN 200910238267A CN 101739827 B CN101739827 B CN 101739827B
- Authority
- CN
- China
- Prior art keywords
- detection algorithm
- pixel
- video image
- target vehicle
- exposure parameter
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000001514 detection method Methods 0.000 claims abstract description 118
- 238000012544 monitoring process Methods 0.000 claims abstract description 17
- 238000012545 processing Methods 0.000 claims abstract description 15
- 230000008859 change Effects 0.000 claims abstract description 12
- 230000001105 regulatory effect Effects 0.000 claims abstract description 7
- 230000008569 process Effects 0.000 claims description 19
- 230000000052 comparative effect Effects 0.000 claims description 8
- 238000010586 diagram Methods 0.000 description 10
- 230000008901 benefit Effects 0.000 description 8
- 238000004040 coloring Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 6
- 230000003044 adaptive effect Effects 0.000 description 5
- 238000000605 extraction Methods 0.000 description 4
- 238000001914 filtration Methods 0.000 description 4
- 230000000877 morphologic effect Effects 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 239000003086 colorant Substances 0.000 description 2
- 238000000691 measurement method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000006116 polymerization reaction Methods 0.000 description 2
- 230000009897 systematic effect Effects 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The embodiment of the invention discloses a vehicle detecting and tracking method which comprises the following steps of: acquiring video images of vehicles, and digitizing and storing the video images; regulating an exposure parameter according to the change of scene brightness, and comparing the relationship of the exposure parameter with the preset exposure threshold; and selecting a detection algorithm according to the comparison result, and processing the stored video images by using the selected detection algorithm to detect and track the object vehicles. The embodiment of the invention improves the vehicle detecting and tracking efficiency and meets the real-time requirement of a vehicle monitoring system. The embodiment of the invention also provides a device using the method.
Description
Technical field
The present invention relates to detection technique field, particularly relate to a kind of vehicle detecting and tracking method and device.
Background technology
As the development trend of 21st century world's road traffic, intelligent transportation system is to apply to whole traffic management system by effectively integrated to advanced infotech, data communication transmission technology, Electronic transducer technology, electron controls technology and computing machine treatment technology etc., set up a kind of on a large scale in, comprehensive playing a role, in real time, multi-transportation and management system accurately and efficiently.Wherein, it is the important component part of intelligent transportation system that the magnitude of traffic flow detects, and can detect in time the vehicle passing through, and counts exactly based on this vehicle flowrate of certain section of road and the movable information of vehicle itself in certain period, as the speed of a motor vehicle, vehicle commander etc.
Traditional magnitude of traffic flow detects the method that adopts buried inductive coil, and owing to existing, construction is complicated, difficult in maintenance, meeting destroys the shortcomings such as road surface, is substituted just gradually by the detection method based on video.In the video of the magnitude of traffic flow detects, according to daytime and night environment difference, the method using and target signature are not identical yet.Wherein, daytime is while carrying out automobile detecting following, image visibility is high, often create road background according to mixed Gauss model, then utilize the methods such as foreground extraction technology, morphologic filtering, profile tracking to detect target area, utilize the gray scale of vehicle or colouring information to carry out target following as feature.Due to road light condition complexity at night, while adopting common CCD (Charge-coupled Device, charge coupled cell) video camera to detect vehicle at night, be difficult to the image of Obtaining Accurate vehicle; The method that some developed countries adopt installation infrared video camera or ccd video camera and thermal camera to combine is obtained road image at night, the method is very effective in the time of detection pedestrian at night, but in the time detecting the vehicle of night running, still can be subject to the impact of high light, ground return light and the ambient light of headlight.In addition, thermal camera expensive, if video camera difference used round the clock can improve hardware cost, the popularization of restriction automobile detecting following system significantly.
At present, prior art adopts common CCD camera to take daytime and nighttime image, and detect vehicle by the headlight in detected image or tail-light night, can obtain good effect.
Inventor is realizing in process of the present invention, and at least there are the following problems to find prior art:
When prior art detects tracking with night to vehicle in the daytime, adopt respectively daytime detection algorithm and night detection algorithm, cannot be according to the variation of external scene change detection algorithm adaptively, reduce the processing speed of vehicle detection and followed the tracks of efficiency, cannot meet the requirement of system real time.
Summary of the invention
The embodiment of the present invention provides a kind of vehicle detecting and tracking method and device, for meeting the requirement of vehicle monitoring system real-time.
The embodiment of the present invention proposes a kind of vehicle detecting and tracking method, comprises the following steps:
The video image of collection vehicle, and described video image is carried out to digitized processing and storage;
Regulate exposure parameter according to the variation of scene brightness, relation between more described exposure parameter and default exposure thresholding, and select detection algorithm according to comparative result, and use the detection algorithm of described selection to process the video image of described storage, target vehicle is carried out to detection and tracking.
Preferably, the described variation according to scene brightness regulates exposure parameter, specifically comprises:
If scene brightness is greater than default luminance threshold, reduce the time shutter; If scene brightness is less than default luminance threshold, increase the time shutter.
Preferably, described detection algorithm comprise daytime detection algorithm and night detection algorithm,
Described according to comparative result selection detection algorithm, specifically comprise:
If described exposure parameter is greater than default exposure thresholding, select detection algorithm at night; If described exposure parameter is less than default exposure thresholding, select detection algorithm in daytime.
Preferably, the detection algorithm of described selection is daytime when detection algorithm,
The detection algorithm that described use is selected carries out detection and tracking to target vehicle, specifically comprises:
Obtain the gray-scale value of each pixel in initial background image according to the video image of described collection, set up background model, and upgrade described background model according to the change of the gray-scale value of described pixel;
Use the background model of described renewal to carry out background difference to the video image of present frame, to detect target vehicle, and the target vehicle detecting is followed the tracks of.
Preferably, the detection algorithm of described selection is night when detection algorithm,
The detection algorithm that described use is selected carries out detection and tracking to target vehicle, specifically comprises:
Detect and follow the tracks of the taillight of described target vehicle, the car light position between predictive frame.
The embodiment of the present invention also proposes a kind of automobile detecting following device, comprising:
Acquisition module, for the video image of collection vehicle, and carries out digitized processing and storage to described video image;
Adjustment module, for regulating exposure parameter according to the variation of scene brightness;
Select module, the relation between the exposure parameter after regulating for more described adjustment module and default exposure thresholding, and select detection algorithm according to comparative result;
Monitoring module, for using the detection algorithm of described selection module selection to process the video image of described acquisition module storage, carries out detection and tracking to target vehicle.
Preferably, described adjustment module, when being greater than default luminance threshold in scene brightness, reduces the time shutter; In the time that scene brightness is less than default luminance threshold, increase the time shutter.
Preferably, described detection algorithm comprise daytime detection algorithm and night detection algorithm,
Described selection module, specifically for being greater than default exposed gate at described exposure parameter in limited time, selects detection algorithm at night; Be less than default exposed gate in limited time at described exposure parameter, select detection algorithm in daytime.
Preferably, the detection algorithm of described selection is daytime when detection algorithm,
Described monitoring module, specifically for obtain the gray-scale value of each pixel in initial background image according to the video image of described collection, sets up background model, and upgrades described background model according to the change of the gray-scale value of described pixel; Use the background model of described renewal to carry out background difference to the video image of present frame, to detect target vehicle, and the target vehicle detecting is followed the tracks of.
Preferably, the detection algorithm of described selection is night when detection algorithm,
Described monitoring module, specifically for detecting and follow the tracks of the taillight of described target vehicle, the car light position between predictive frame.
The technical scheme of the embodiment of the present invention has the following advantages, because make full use of the relation between exposure parameter and the external scene of video camera, select adopt daytime detection algorithm or night detection algorithm, having solved original supervisory system cannot adaptive shortcoming, has effectively met the demand of round-the-clock monitoring, has avoided complicated scene judgement, processing speed is higher, can requirement of real time, improve and follow the tracks of efficiency, be convenient to be applied in real-time follow-up scene.
Brief description of the drawings
In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, to the accompanying drawing of required use in embodiment or description of the Prior Art be briefly described below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, do not paying under the prerequisite of creative work, can also obtain according to these accompanying drawings other accompanying drawing.
Fig. 1 is a kind of vehicle detecting and tracking method process flow diagram in the embodiment of the present invention one;
Fig. 2 is a kind of vehicle detecting and tracking method process flow diagram in the embodiment of the present invention two;
Fig. 3 is vehicle detecting and tracking method process flow diagram under scene in daytime in the embodiment of the present invention three;
Fig. 4 is vehicle detecting and tracking method process flow diagram under scene at night in the embodiment of the present invention four;
Fig. 5 is a kind of automobile detecting following apparatus structure schematic diagram in the embodiment of the present invention five.
Embodiment
The embodiment of the present invention provides a kind of adaptive vehicle checking method and device round the clock based on video camera exposure parameter, regulate adaptively exposure parameter according to the variation of external scene light intensity, and a reasonably exposure thresholding is set, effectively judge on-the-spot external scene situation according to the relation of exposure thresholding and exposure parameter, select detection algorithm to carry out the detection and tracking of vehicle.Wherein, night, detection algorithm was by analyzing R in tail-light, G, tri-color components of B, can in coloured image, detect tail-light at night, and by tail-light is carried out to continuous tracking, obtain the movable information of tail-light, again in conjunction with having the prioris such as fixing position relationship between tail-light on same car, can mate the tail-light in image, the tail-light that belongs to same car is returned together, take full advantage of colouring information, do not need to calculate morphological operator, reduce calculated amount, detection speed is obviously improved, can requirement of real time.The present invention also adopts symmetry Measurement Method as judgement foundation, and the degree of polymerization of the ordinate between the taillight that judgement detects, carries out the pairing of car light.
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is clearly and completely described, obviously, described embodiment is the present invention's part embodiment, instead of whole embodiment.Based on the embodiment in the present invention, those of ordinary skill in the art, not making the every other embodiment obtaining under creative work prerequisite, belong to the scope of protection of the invention.
As shown in Figure 1, be a kind of vehicle detecting and tracking method process flow diagram in the embodiment of the present invention one, comprise the following steps:
Step 101, the video image of collection vehicle, and this video image is carried out to digitized processing and storage.
Step 102, regulate exposure parameter according to the variation of scene brightness, the relatively relation between this exposure parameter and default exposure thresholding, and according to comparative result selection detection algorithm, use the detection algorithm of selecting to process the video image of storage, target vehicle is carried out to detection and tracking.
Particularly, if scene brightness is greater than default luminance threshold, reduce the time shutter; If scene brightness is less than default luminance threshold, increase the time shutter.If exposure parameter is greater than default exposure thresholding, select night detection algorithm to carry out detection and tracking to target vehicle; If exposure parameter is less than default exposure thresholding, select daytime detection algorithm to carry out detection and tracking to target vehicle.
The technical scheme of the embodiment of the present invention has the following advantages, because make full use of the relation between exposure parameter and the external scene of video camera, select adopt daytime detection algorithm or night detection algorithm, having solved original supervisory system cannot adaptive shortcoming, has effectively met the demand of round-the-clock monitoring, has avoided complicated scene judgement, processing speed is higher, can requirement of real time, improve and follow the tracks of efficiency, be convenient to be applied in real-time follow-up scene.
As shown in Figure 2, be a kind of vehicle detecting and tracking method process flow diagram in the embodiment of the present invention two, comprise the following steps:
Step 201, the video image of collection vehicle, and this video image is carried out to digitized processing and storage.
Particularly, can utilize the video image of camera collection vehicle and this video image is carried out to digitizing, the video image of driving vehicle on the section of ccd video camera picked-up is got off by sequence continuous capturing and deposit in frame memory after A/D (Analog to Digital, analog quantity is to digital quantity) conversion, digitizing by image capturing system.Gather before video image, also need by system initialization, systematic parameter to be set, this systematic parameter comprises the frame number of collection image per second and the threshold value of image binaryzation etc.
Step 202, carries out pre-service to the video image collecting.
Particularly, there is the problems such as fuzzy and noise due to the video image of dynamic acquisition, need to carry out the pre-service such as filtering, noise reduction, greyscale transformation and binaryzation.
Step 203, in the time that scene brightness is greater than default luminance threshold, reduces the time shutter; In the time that scene brightness is less than default luminance threshold, increase the time shutter.
Wherein, corresponding to daytime and night, the exposure parameter of video camera is also different.
Step 204, relatively the relation between exposure parameter and default exposure thresholding, if exposure parameter is greater than default exposure thresholding, performs step 205; If exposure parameter is less than default exposure thresholding, perform step 206.
Step 205, selects detection algorithm at night.
Particularly, adopt taillight technology owing to detecting night, do not have headlight direct projection camera and cause the conversion that parameter is violent.Simultaneously, for fear of noise and abnormal interference, state toggle count device can be set, when scene is switched to night from daytime, by daytime detection algorithm be switched to night when detection algorithm, carry out timing by state toggle count device, after the video image collecting continues N frame and scene and do not switch, confirm the detection algorithm of selecting.
Step 206, selects detection algorithm in daytime.
Particularly, when scene is switched to daytime from night, by night detection algorithm be switched to daytime when detection algorithm, carry out timing by state toggle count device, after the video image collecting continues N frame and scene and do not switch, confirm the detection algorithm of selecting.
Step 207, is used the detection algorithm of selecting to process the video image of storage, and target vehicle is carried out to detection and tracking.
It should be noted that, the inventive method can be adjusted each sequence of steps according to actual needs.
The technical scheme of the embodiment of the present invention has the following advantages, because make full use of the relation between exposure parameter and the external scene of video camera, select adopt daytime detection algorithm or night detection algorithm, having solved original supervisory system cannot adaptive shortcoming, effectively meet the demand of round-the-clock monitoring, avoid complicated scene judgement, processing speed is higher, can requirement of real time, improve and follow the tracks of efficiency, can be used for the target following under real-time conditions, efficiency is high, good stability, convenient embedding in front end system, saves hardware cost.
As shown in Figure 3, be vehicle detecting and tracking method process flow diagram under scene in the daytime in the embodiment of the present invention three, the detection algorithm of selection is detection algorithm in daytime, the method comprises the following steps:
Step 301, obtains the gray-scale value of each pixel in initial background image according to the video image gathering, and sets up background model.
Particularly, can be according to draw each pixel gray-scale value initial background image from the some frame video images that collect.In the process of setting up background model, can be in the situation that scene have moving target background extraction image, but in actual applications, the monitoring of most of scenes, the monitoring of for example road, all be difficult to ensure not have in video scene vehicle or the object appearance of motion, therefore cannot meet scene does not have the requirement of moving target, and need to adopt a kind of method of background extraction image can there is moving target in scene in the situation that.
The embodiment of the present invention adopts the modeling method based on Gaussian statistics model, in video sequence image, the gray-scale value of each pixel meets Gaussian distribution, average and the mean square deviation of (x, y) point of n two field picture before therefore can calculating by following formula:
Wherein, μ
n(x, y) and σ
n(x, y) represents respectively average and the mean square deviation of (x, y) point of front n two field picture.
Further, can utilize limited two field picture by following Formula background model:
Wherein, f
ik(x, y) meets | f
ik(x, y)-μ
n(x, y) |≤β σ
n(x, y), β is predefined constant, when frame number is more, can be set to 1.
Step 302, upgrades background model according to the change of the gray-scale value of pixel.
Particularly, along with the change of the environmental factors such as intensity of illumination, background model must upgrade in time.The embodiment of the present invention adopts following update strategy, first utilizes present frame and former frame to carry out frame poor, in the time that two frames differ that the gray-scale value of certain pixel is greater than certain threshold value afterwards, judges that this pixel is sport foreground region point; Otherwise, judge that this pixel is background area point.
Further, sport foreground region point is retained, adopts formula below to upgrade to background area point:
B
k(x，y)＝αB
k-1(x，y)+(1-α)f
k(x，y) (3-4)
Wherein, B
k(x, y) is the background model in k frame moment; α is renewal rate, for representing the renewal speed of background model, and 0≤α≤1.
Step 303, is used the background model of upgrading to carry out background difference to the video image of present frame, to detect target vehicle.
Particularly, use the background model B upgrading
k(x, y) carries out background difference to present frame, and the picture point that is greater than a certain threshold value T is the point of moving target, and concrete formula is as follows:
Wherein, M is the binary map matrix that moving region is cut apart, and M (x, y)=1 represents sport foreground region, and M (x, y)=0 represents background area.
Step 304, follows the tracks of the target vehicle detecting.
Particularly, can utilize color or the edge feature of target vehicle, adopt the trackings such as Meanshift to follow the tracks of target vehicle.
It should be noted that, the inventive method can be adjusted each sequence of steps according to actual needs.
The technical scheme of the embodiment of the present invention has the following advantages, and daytime, detection algorithm algorithm made full use of in video sequence image, and the gray-scale value of each pixel meets the feature of Gaussian distribution and carries out Gauss's modeling, then carry out foreground extraction, complete the detection of vehicle, method simple practical, real-time is good.
As shown in Figure 4, be vehicle detecting and tracking method process flow diagram under scene at the night in the embodiment of the present invention four, the detection algorithm of selection is detection algorithm at night, the method comprises the following steps:
Step 401, the taillight of detection target vehicle.
Particularly, the embodiment of the present invention is used CCD colour TV camera, in the coloured image obtaining, shades of colour is by RGB (Red Green Blue, RGB) three primary colours formation, due to night light darker, the colouring information of car body can not accurately obtain, but the colouring information of red tail-light can be caught by video camera exactly.According to three primary colours theory, by analyzing the color component separately of R, G, B in tail-light, can in coloured image, detect tail-light at night, and the other light sources such as tail-light, traffic lights and street lamp are made a distinction.
Because the tail-light quantity of various vehicles on road is different, therefore, can not be simply by tail-light quantity be carried out to statistical vehicle flowrate divided by 2.By tail-light is carried out to continuous tracking, obtain the movable information of tail-light, again in conjunction with having the prioris such as fixing position relationship between tail-light on same car, can mate the tail-light in image, the tail-light that belongs to same car is returned together, can count exactly vehicle flowrate.
In above-mentioned testing process, video camera is that the direction of sailing out of towards vehicle is taken, and image is subject to the reflective impact of headstock high light and ground very little, and take full advantage of colouring information, and do not need to calculate morphological operator, reduce calculated amount, detection speed is obviously improved, can requirement of real time.
By known to the analysis of nighttime image color histogram, utilize the feature of the corresponding point gray-scale value of R, G, tri-color channels of B, tail-light can be separated from nighttime image.The color space vector of supposing pixel i in image is { R
i, G
i, B
i, define the partially red horizontal r of pixel i
ifor:
Wherein, r
iless, more deflection is red to show pixel i, if pixel i belongs to tail-light, and r
i∈ [0,1].
Although utilize the accurately partially red level of every in computed image of formula (4-1), to judge whether this point belongs to tail-light, and operand is larger.In practical application, only need to consider the partially red level in a certain region in image, just can determine the position of tail-light, the partially red horizontal r of defined range S
sfor:
Utilize formula (4-2), for possible tail-light region, only need carry out a multiplication and a division arithmetic, obviously in comparison chart picture, institute is a little all directly little with the operand of the partially red level of use formula (4-1) calculating.
The taillight algorithm of target vehicle specifically comprises the following steps:
The R passage of each pixel in A, detection nighttime image, if R
i>=R
th, retain this picture point color space vector; Otherwise, this color space vector is set to 0.Wherein, I is that original image is, R
thfor R passage thresholding, through processing, in image, partially red hot spot and white will be only left.
The partially red horizontal r of spot area in B, use formula (4-2) calculation procedure A image after treatment
sif, r
s>=r
th, retain this picture point color space vector; Otherwise, this color space vector is set to 0.Wherein, r
thfor the partially red thresholding of color, through processing, in image, be only left real tail-light region and possible inverted image thereof.
C, the pairing of interframe car light.
For light for vehicle, under normal circumstances, the position of both sides taillight in short transverse is consistent, and two taillights cross central point line and horizontal direction angle should be very little, and perfect condition is to be 0.In addition, both sides taillight has necessarily symmetric in the horizontal direction.Based on above analysis, the embodiment of the present invention adopts symmetry Measurement Method as judgement foundation, the degree of polymerization of the ordinate between the taillight that judgement detects, then horizontal direction rises, from left to right, two two pairings, are combined into a vehicle target and obtain its center { C
xi, C
yi, i=1 .., n.
Step 402, the taillight of tracking target vehicle.
Wherein, the tracking of moving target is a complicated process, first needs new moving target to carry out initialization, obtains the initialization kinematic parameter of moving target, also will in next frame, search for optimum matching to initialized moving target simultaneously.Too simplify owing to finding coupling using moving target as point, easily cause the judgement of many mistakes, therefore, the embodiment of the present invention represents moving region with boundary rectangle, be equivalent to boundary rectangle to mate tracking, after moving target symbolism, make judgement more accurate, result is more credible.
The embodiment of the present invention is used the most adjacent method of approximation to carry out association to vehicle sequence and follows the tracks of, and information easy to use and need to be known is few.Owing to being unable to estimate the kinematic parameter of emerging target, therefore, directly according to the propinquity spatially of each target in this target and former frame object chain and the difference degree of size, in whole frame, look for best coupling, if matching degree exceedes certain threshold value,, using this matching relationship as this continuity of motion rectangle in next frame, be recorded as identical motion label; Otherwise, confirm as fresh target.Meanwhile, detect the target that does not obtain coupling in former frame object chain, adopt position prediction module to predict, if exceed image border, be considered as target and disappear; Otherwise, predict, and write in the object chain of present frame, and indicate that this parameter is for predicting the outcome, if continuous three frames all belong to and predict the outcome, be judged to be track rejection.
Step 403, the car light position between predictive frame.
Particularly, because the time that hot-short occurs in the visual field arranging is very short, be equivalent to the time of wave filter initialization needs, therefore, the embodiment of the present invention does not adopt traditional alpha-beta filtering or Kalman filtering.In addition, easily estimate because vehicle interframe moves maximum magnitude Scale, suppose that video frame rate f is 25, the upper limit maxV of road speed limit is 144km/h, i.e. 40m/s, and the minimum vehicle commander L of vehicle is 2m,
From above inference, next frame car light is being put forward direction L distance to position is inevitable when front car light contraposition, and at the interior car light that can not have other vehicles of this distance.So right tracking is simple for car light, as long as using above-mentioned distance as target prediction scope.In addition, could be detected by algorithm because vehicle must at least exist three, four frames in video, formula (4-3) gives the minimum length that the visual field arranges.
It should be noted that, the inventive method can be adjusted each sequence of steps according to actual needs.
The technical scheme of the embodiment of the present invention has the following advantages, and night, detection algorithm made full use of colouring information and the movable information of car light, adopts space symmetr strategy to carry out taillight pairing, identifies single target vehicle; Meanwhile, be the situation of avoiding light for vehicle to block, introduce the function of position prediction, effectively avoid track rejection, solve the many and not high problem of accuracy consuming time while calculating morphological operator in traditional algorithm.
As shown in Figure 5, be a kind of automobile detecting following apparatus structure schematic diagram in the embodiment of the present invention five, comprising:
Acquisition module 510, for the video image of collection vehicle, and carries out digitized processing and storage to described video image.
Adjustment module 520, for regulating exposure parameter according to the variation of scene brightness.
Particularly, above-mentioned adjustment module 520, when being greater than default luminance threshold in scene brightness, reduces the time shutter; In the time that scene brightness is less than default luminance threshold, increase the time shutter.
Select module 530, the relation between the exposure parameter after regulating for comparison adjustment module 520 and default exposure thresholding, and select detection algorithm according to comparative result.
Wherein, above-mentioned detection algorithm comprise daytime detection algorithm and night detection algorithm, above-mentioned selection module 530, specifically for being greater than default exposed gate at described exposure parameter in limited time, selects detection algorithm at night; Be less than default exposed gate in limited time at described exposure parameter, select detection algorithm in daytime.
Monitoring module 540, selects the detection algorithm that module 530 is selected to process the video image that acquisition module 510 is stored for using, and target vehicle is carried out to detection and tracking.
Particularly, the detection algorithm of above-mentioned selection is that daytime is when detection algorithm, above-mentioned monitoring module 540, specifically for obtain the gray-scale value of each pixel in initial background image according to the video image of described collection, set up background model, and upgrade described background model according to the change of the gray-scale value of described pixel; Use the background model of described renewal to carry out background difference to the video image of present frame, to detect target vehicle, and the target vehicle detecting is followed the tracks of.
The detection algorithm of above-mentioned selection is night when detection algorithm, above-mentioned monitoring module 540, and specifically for detecting and follow the tracks of the taillight of described target vehicle, the car light position between predictive frame.
The technical scheme of the embodiment of the present invention has the following advantages, because make full use of the relation between exposure parameter and the external scene of video camera, select adopt daytime detection algorithm or night detection algorithm, having solved original supervisory system cannot adaptive shortcoming, has effectively met the demand of round-the-clock monitoring, has avoided complicated scene judgement, processing speed is higher, can requirement of real time, improve and follow the tracks of efficiency, can be used for the target following under real-time conditions.Meanwhile, apparatus module clear in structure, the each several part division of labor is clear and definite, and independence is strong, has improved overall stability and reliability.
Through the above description of the embodiments, those skilled in the art can be well understood to the mode that the present invention can add essential general hardware platform by software and realize, and can certainly pass through hardware, but in a lot of situation, the former is better embodiment.Based on such understanding, the part that technical scheme of the present invention contributes to prior art in essence in other words can embody with the form of software product, this computer software product is stored in a storage medium, comprise that some instructions (can be mobile phones in order to make a station terminal equipment, personal computer, server, or the network equipment etc.) carry out the method described in each embodiment of the present invention.
The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the premise without departing from the principles of the invention; can also make some improvements and modifications, these improvements and modifications also should be looked protection scope of the present invention.
It will be appreciated by those skilled in the art that the module in the device in embodiment can be distributed in the device of embodiment according to embodiment description, also can carry out respective change and be arranged in the one or more devices that are different from the present embodiment.The module of above-described embodiment can be integrated in one, and also can separate deployment, can merge into a module, also can further split into multiple submodules.
The invention described above embodiment sequence number, just to describing, does not represent the quality of embodiment.
Disclosed is above only several specific embodiment of the present invention, and still, the present invention is not limited thereto, and the changes that any person skilled in the art can think of all should fall into protection scope of the present invention.
Claims (4)
1. a vehicle detecting and tracking method, is characterized in that, comprises the following steps:
The video image of collection vehicle, and described video image is carried out to digitized processing and storage;
Regulate exposure parameter according to the variation of scene brightness, relation between more described exposure parameter and default exposure thresholding, and select detection algorithm according to comparative result, and use the detection algorithm of described selection to process the video image of described storage, target vehicle is carried out to detection and tracking;
Wherein, described detection algorithm comprise daytime detection algorithm and night detection algorithm, described according to comparative result select detection algorithm, specifically comprise: if described exposure parameter is greater than default exposure thresholding, select detection algorithm at night; If described exposure parameter is less than default exposure thresholding, select detection algorithm in daytime;
The detection algorithm of described selection is that daytime is when detection algorithm, the detection algorithm that described use is selected carries out detection and tracking to target vehicle, specifically comprise: the gray-scale value that obtains each pixel in initial background image according to the video image of described collection, set up background model, and upgrade described background model according to the change of the gray-scale value of described pixel; Use the background model of described renewal to carry out background difference to the video image of present frame, to detect target vehicle, and the target vehicle detecting is followed the tracks of;
Upgrade background model according to the change of the gray-scale value of pixel, be specially: utilize present frame and former frame to carry out frame poor, in the time that two frames differ that the gray-scale value of certain pixel is greater than certain threshold value afterwards, judge that this pixel is sport foreground region point; Otherwise, judge that this pixel is background area point;
The described variation according to scene brightness regulates exposure parameter, specifically comprises:
If scene brightness is greater than default luminance threshold, reduce the time shutter; If scene brightness is less than default luminance threshold, increase the time shutter.
2. the method for claim 1, is characterized in that, the detection algorithm of described selection is night when detection algorithm,
The detection algorithm that described use is selected carries out detection and tracking to target vehicle, specifically comprises:
Detect and follow the tracks of the taillight of described target vehicle, the car light position between predictive frame.
3. an automobile detecting following device, is characterized in that, comprising:
Acquisition module, for the video image of collection vehicle, and carries out digitized processing and storage to described video image;
Adjustment module, for regulating exposure parameter according to the variation of scene brightness;
Select module, the relation between the exposure parameter after regulating for more described adjustment module and default exposure thresholding, and select detection algorithm according to comparative result;
Monitoring module, for using the detection algorithm of described selection module selection to process the video image of described acquisition module storage, carries out detection and tracking to target vehicle;
Wherein, described detection algorithm comprise daytime detection algorithm and night detection algorithm, described selection module, specifically for being greater than default exposed gate at described exposure parameter in limited time, selects detection algorithm at night; Be less than default exposed gate in limited time at described exposure parameter, select detection algorithm in daytime;
Described monitoring module, specifically for, when the detection algorithm of described selection is daytime when detection algorithm, obtain the gray-scale value of each pixel in initial background image according to the video image of described collection, set up background model, and upgrade described background model according to the change of the gray-scale value of described pixel; Use the background model of described renewal to carry out background difference to the video image of present frame, to detect target vehicle, and the target vehicle detecting is followed the tracks of;
Upgrade background model according to the change of the gray-scale value of pixel, be specially: utilize present frame and former frame to carry out frame poor, in the time that two frames differ that the gray-scale value of certain pixel is greater than certain threshold value afterwards, judge that this pixel is sport foreground region point; Otherwise, judge that this pixel is background area point;
Described adjustment module, when being greater than default luminance threshold in scene brightness, reduces the time shutter; In the time that scene brightness is less than default luminance threshold, increase the time shutter.
4. device as claimed in claim 3, is characterized in that, the detection algorithm of described selection is night when detection algorithm,
Described monitoring module, specifically for detecting and follow the tracks of the taillight of described target vehicle, the car light position between predictive frame.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910238267.3A CN101739827B (en) | 2009-11-24 | 2009-11-24 | Vehicle detecting and tracking method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910238267.3A CN101739827B (en) | 2009-11-24 | 2009-11-24 | Vehicle detecting and tracking method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101739827A CN101739827A (en) | 2010-06-16 |
| CN101739827B true CN101739827B (en) | 2014-10-29 |
Family
ID=42463256
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200910238267.3A Active CN101739827B (en) | 2009-11-24 | 2009-11-24 | Vehicle detecting and tracking method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101739827B (en) |
Families Citing this family (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103177237A (en) * | 2011-12-22 | 2013-06-26 | 中国移动通信集团河北有限公司 | Video monitoring method and device based on on-line lasers |
| CN103106796A (en) * | 2013-01-15 | 2013-05-15 | 江苏大学 | Vehicle detection method and device of intelligent traffic surveillance and control system |
| CN103489317A (en) * | 2013-10-10 | 2014-01-01 | 扬州瑞控汽车电子有限公司 | Method for detecting vehicle in different scenes |
| CN103886292B (en) * | 2014-03-20 | 2017-02-08 | 杭州电子科技大学 | Night vehicle target stable tracking method based on machine vision |
| CN104021679A (en) * | 2014-05-30 | 2014-09-03 | 宁波海视智能系统有限公司 | Traffic flow statistic method based on video |
| CN104200197A (en) * | 2014-08-18 | 2014-12-10 | 北京邮电大学 | Three-dimensional human body behavior recognition method and device |
| CN108711283A (en) * | 2015-02-10 | 2018-10-26 | 杭州海存信息技术有限公司 | The night monitoring to park cars |
| CN105718893B (en) * | 2016-01-22 | 2019-01-08 | 江苏大学 | A kind of light for vehicle for night-environment is to detection method |
| TWI618647B (en) * | 2016-02-02 | 2018-03-21 | 財團法人資訊工業策進會 | System and method of detection, tracking and identification of evolutionary adaptation of vehicle lamp |
| CN106163064A (en) * | 2016-09-09 | 2016-11-23 | 成都通甲优博科技有限责任公司 | Road lamp energy-saving automatic control system based on road target detection and method |
| CN106874869B (en) * | 2017-02-15 | 2018-07-27 | 北京航空航天大学 | A kind of information safety system and method for vehicle-mounted visual perception |
| CN107249104A (en) * | 2017-06-15 | 2017-10-13 | 武汉云衡智能科技有限公司 | A kind of drive recorder smart camera automatic explosion method |
| CN107516417B (en) * | 2017-08-21 | 2019-09-17 | 中国科学院软件研究所 | A kind of real-time highway flow estimation method for excavating spatial and temporal association |
| CN108538052A (en) * | 2018-03-05 | 2018-09-14 | 华南理工大学 | Night traffic flow rate testing methods based on headlight track following and dynamic pairing |
| CN108764102A (en) * | 2018-05-22 | 2018-11-06 | 哈尔滨理工大学 | A kind of characteristic recognition method of the front vehicles image based on color extraction algorithm |
| CN110084111B (en) * | 2019-03-19 | 2023-08-25 | 江苏大学 | Rapid night vehicle detection method applied to self-adaptive high beam |
| CN111507237A (en) * | 2020-04-13 | 2020-08-07 | 北京巨视科技有限公司 | Video-based abnormal event detection method and device applied to expressway |
| CN111966857B (en) * | 2020-08-19 | 2023-09-29 | 南京英德利汽车有限公司 | Method and system for detecting refitted vehicle |
| CN116878402B (en) * | 2023-07-11 | 2024-07-19 | 北京博科测试系统股份有限公司 | Non-contact wheel arch measuring sensor and method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101017573A (en) * | 2007-02-09 | 2007-08-15 | 南京大学 | Method for detecting and identifying moving target based on video monitoring |
-
2009
- 2009-11-24 CN CN200910238267.3A patent/CN101739827B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101017573A (en) * | 2007-02-09 | 2007-08-15 | 南京大学 | Method for detecting and identifying moving target based on video monitoring |
Non-Patent Citations (1)
| Title |
|---|
| 吴海涛.复杂环境下的夜间车辆检测与交通视频昼夜亮度变化模型研究.《中国优秀硕士学位论文全文数据库（电子期刊）信息科技辑》.2007,(第6期),第15-50页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101739827A (en) | 2010-06-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101739827B (en) | Vehicle detecting and tracking method and device | |
| CN101727748B (en) | Method, system and equipment for monitoring vehicles based on vehicle taillight detection | |
| CN101656023B (en) | Management method of indoor car park in video monitor mode | |
| US9704060B2 (en) | Method for detecting traffic violation | |
| CN101872546B (en) | Video-based method for rapidly detecting transit vehicles | |
| CN105913685A (en) | Video surveillance-based carport recognition and intelligent guide method | |
| CN107808392B (en) | Automatic tracking and positioning method and system for security check vehicle in open scene | |
| US9672434B2 (en) | Video-based system and method for parking occupancy detection | |
| CN103069434B (en) | For the method and system of multi-mode video case index | |
| CN101827204B (en) | Method and system for detecting moving object | |
| CN101510356B (en) | Video detection system and data processing device thereof, video detection method | |
| CN102385803B (en) | All-weather urban vehicle tracking and counting method based on video monitoring | |
| US8019157B2 (en) | Method of vehicle segmentation and counting for nighttime video frames | |
| CN110069986A (en) | A kind of traffic lights recognition methods and system based on mixed model | |
| CN105160691A (en) | Color histogram based vehicle body color identification method | |
| CN104601964A (en) | Non-overlap vision field trans-camera indoor pedestrian target tracking method and non-overlap vision field trans-camera indoor pedestrian target tracking system | |
| CN103871079A (en) | Vehicle tracking method based on machine learning and optical flow | |
| EP2013817A2 (en) | Video segmentation using statistical pixel modeling | |
| CN108320575A (en) | A kind of guiding of parking stall and reverse vehicle searching system and its parking space-detecting device, method | |
| CN112017445B (en) | Pedestrian violation prediction and motion trail tracking system and method | |
| CN102496281A (en) | Vehicle red-light violation detection method based on combination of tracking and virtual loop | |
| CN102789686B (en) | Road traffic flow detecting method based on road surface brightness composite mode recognition | |
| CN104751138B (en) | A kind of vehicle mounted infrared image colorization DAS (Driver Assistant System) | |
| Wang et al. | Video image vehicle detection system for signaled traffic intersection | |
| CN108109414A (en) | A kind of preventing road monitoring system and method based on vehicle assistant drive |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210207 Address after: No. 607, 6th floor, shining building, 35 Xueyuan Road, Haidian District, Beijing 100083 Patentee after: BEIJING VIMICRO ARTIFICIAL INTELLIGENCE CHIP TECHNOLOGY Co.,Ltd. Address before: 100083, Haidian District, Xueyuan Road, Beijing No. 35, Nanjing Ning building, 16 floor Patentee before: Vimicro Corp. |