CN103258432A - Traffic accident automatic identification processing method and system based on videos - Google Patents
Traffic accident automatic identification processing method and system based on videos Download PDFInfo
- Publication number
- CN103258432A CN103258432A CN2013101395456A CN201310139545A CN103258432A CN 103258432 A CN103258432 A CN 103258432A CN 2013101395456 A CN2013101395456 A CN 2013101395456A CN 201310139545 A CN201310139545 A CN 201310139545A CN 103258432 A CN103258432 A CN 103258432A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- traffic accident
- accident
- traffic
- video
- 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.)
- Granted
Links
Images
Abstract
The invention discloses a traffic accident automatic identification processing method and system based on videos. The traffic accident automatic identification processing method and system based on the videos comprises the steps of respectively setting up three-dimensional feature libraries of different vehicle types, obtaining road video image sequences, carrying out foreground vehicle separation based on background modeling, carrying out static target judgment based on target centroid displacement, recognizing the vehicle types based on a vehicle outer contour recognition algorithm of three-dimensional modeling, judging whether a traffic accident occurs or not according to whether contour deformation occurs or not, extracting improved SIFT feature points of the vehicles based on an improved SIFT feature recognition algorithm of vehicles of the three-dimensional modeling, and judging whether a traffic accident occurs or not according to comparison results of the improved SIFT feature points. The traffic accident automatic identification processing method and system based on the videos judges whether vehicles on a current road are driven safely or not by analyzing images collected by a camera, collects information of a traffic accident scene in first time of a traffic accident, and transfers the information to a commanding monitoring centre. Workers can effectively and timely work by watching the videos, and the traffic accident automatic identification processing method and system based on the videos plays a huge role in perfecting a whole intelligent traffic system.
Description
Technical field
The present invention relates to intelligent transportation video frequency graphic monitoring and video image analysis field, particularly disposal route and system are identified in a kind of traffic hazard based on video automatically.
Background technology
In recent years, along with the continuous progress of fast development of national economy and society, road extends in all direction, and the quantity of automobile grows with each passing day especially, and thing followed traffic problems are also increasingly serious, traffic jam.How to realize real-time monitoring, scheduling and the control of traffic scene, set up focus and current problem demanding prompt solution that effective intelligent transportation system becomes domestic and international concern already.Under this background, provide a kind of more real-time, accurately and high-efficiency method based on the visual intelligent treatment technology of computer vision and Digital Image Processing to the processing of traffic hazard, and provide technical support to the rescue of traffic hazard, the rehabilitations such as confirmation of responsibility of accident.
At present, domestic less at the research of the automatic identification of traffic hazard based on video, research to traffic hazard mainly comprises following several prior art: 1, toroidal inductor method, this method and technology maturation, not influenced by the weather political reform, but this technology is mainly carried out the estimation of traffic accident based on travel condition of vehicle; 2, it is more that infrared detection method, the method obtain parameter, but infrared being subject to disturb, and whether can't distinguish by thing is vehicle; 3, ultrasonic Detection Method, the method directionality is good, reflection potential is strong, but to obtain parameter less, is subjected to temperature, climate effect big.And at present based on the traffic hazard identification of video technique, only utilize simple car speed parameter, and can not well identify traffic hazard, can not accomplish Realtime Alerts.
Summary of the invention
The purpose of this invention is to provide a kind of traffic hazard based on video and identify disposal route and system automatically, be used for to realize detecting traffic hazard fast and effectively, and Realtime Alerts and to effective management of accident.
To achieve these goals, the present invention adopts following technical scheme:
Disposal route is identified in traffic hazard based on video automatically, it is characterized in that, specifically may further comprise the steps:
Step S10: different automobile types is set up its three-dimensional feature storehouse respectively;
Step S11: obtain the road sequence of video images;
Step S12: the prospect vehicle based on background modeling separates;
Step S13: the static target of based target barycenter displacement is judged;
Step S14: based on the vehicle outline recognizer of three-dimensional modeling, identify vehicle, according to profile whether deformation takes place again and judge whether to get into an accident, as judge to get into an accident and then report to the police, otherwise enter step S15, judge by improving the SIFT unique point;
Step S15: based on the improvement SIFT feature recognition algorithms of the vehicle of three-dimensional modeling: extract the improvement SIFT unique point of vehicle, mate with local feature point in the vehicle three-dimensional feature storehouse, and comprehensively judge whether to get into an accident according to matching result.
The present invention further improves and is: among the described step S10 different automobile types is set up its three-dimensional feature storehouse step respectively and comprise:
S101) obtain different automobile types at the image of the different attitude angles of three dimensions;
S102) extract the unique point of vehicle outline in the image;
S103) extract the improvement SIFT unique point in the vehicle closed curve that outline surrounds in the image;
S104) use the above-mentioned two category features point that extracts to set up vehicle three-dimensional feature storehouse.
The present invention further improves and is: obtaining the road sequence of video images among the described step S11 is to obtain the road video image by setting up single fixed cameras with a certain interval at road.
The present invention further improves and is: prospect vehicle separating step comprises among the described step S12:
S121) for colour imagery shot, adopt the background modeling method based on color space cluster road model; For the black and white camera, adopt the background modeling method based on mixed Gauss model;
S122) obtain prospect and background image based on the image difference point-score;
S123) utilize morphology that the vehicle segmentation result is improved.
The present invention further improves and is: the stationary vehicle identification step based on image difference among the described step S13 comprises:
S131) utilize 8 neighborhood communicating methods that moving target is carried out mark respectively;
S132) moving target that marks is calculated the size of its barycenter and connected region respectively;
S133) compare with the former frame image, judged whether that target is offset.
The present invention further improves and is: the vehicle outline recognizer step based on three-dimensional modeling among the described step S14 comprises:
S141) use the canny operator to obtain the outline of prospect vehicle;
S142) outline that obtains is carried out mathematical description, obtain its mathematic(al) representation;
S143) angle point in the contouring curve mates with the three-dimensional vehicle ' s contour storehouse of prior foundation as unique point;
S144) identify vehicle model, and judge whether the vehicle outline changes;
S145) change as outline, carry out traffic accident and report to the police, judge otherwise improve the SIFT feature by step S15.
The present invention further improves and is: the vehicle improvement SIFT feature recognition algorithms step based on three-dimensional modeling among the described step S15 comprises:
S151) obtain the improvement SIFT feature of vehicle;
S152) mate with the three-dimensional feature storehouse of setting up in advance.
The automatic identification processing system of a kind of traffic hazard based on video comprises communication subsystem, information storage subsystem, accident recognition subsystem and Incident Management subsystem;
Wherein, described communication subsystem, 1) be used for the video sequence of camera acquisition is transferred to workstation; 2) be used for containing the accident transmission of video images to the server of district commander Surveillance center haveing through the workstation processing and identification; 3) finish communication between district, city commander Surveillance center and the server;
Described information storage subsystem, the original video of storage traffic accident period of right time, the identifying information that corresponding video reflects the traffic accident grade, traffic accident be dot information and relevant rehabilitation information relatively;
Described accident recognition subsystem, 1) the identification traffic accident takes place, and intercepting comprises the video information of traffic accident generating process; 2) identify intensity grade according to relevant specification or regulation;
Described Incident Management subsystem is used for 1) the traffic accident warning; 2) traffic accident ranking compositor; 3) accident rehabilitation; 4) historical information inquiry, statistics.
The present invention further improves and is: described accident recognition subsystem comprises: background modeling module, prospect vehicle extraction module, vehicle ' s contour identification module and vehicle improve SIFT feature identification module;
Background modeling module: utilize the road video image information, adopt respectively based on the method for color space cluster with based on gauss hybrid models for colored and black and white camera and carry out background modeling;
Prospect vehicle extraction module: utilize the background difference algorithm, present frame and background are done difference, obtain the prospect vehicle, utilize eight connected domain algorithms that moving target is separately identified;
Vehicle ' s contour identification module: use the canny operator to obtain the profile of prospect vehicle, and mate with the three-dimensional vehicle ' s contour storehouse of prior foundation, at first identify vehicle, then if vehicle ' s contour deformation is obvious, be judged to be traffic accident, otherwise use local SFIT feature to differentiate;
Vehicle improves SIFT feature identification module: extract vehicle and improve the SIFT feature, mate with the feature database of prior foundation, judge whether to be traffic accident.
Compared with prior art, the invention has the beneficial effects as follows: by analyze camera collection to image judge whether safety traffic of vehicle on the current road surface, and the information of traffic accident scene under gathering in the very first time that traffic accident takes place transfers to commander Surveillance center.The staff of commander Surveillance center can carry out the work timely and effectively by checking video, has so both saved manpower, can obtain how effective conclusion again, for improving whole intelligent transportation system very big effect is arranged.
Description of drawings
Fig. 1 is that process flow figure is identified in the traffic hazard that the present invention is based on video automatically;
Fig. 2 is the detailed process process flow diagram of different automobile types being set up its three-dimensional feature storehouse respectively;
Figure 3 shows that the prospect vehicle based on background modeling separates the detailed process process flow diagram;
Figure 4 shows that the vehicle outline recognizer particular flow sheet based on three-dimensional modeling;
Figure 5 shows that the improvement SIFT feature recognition algorithms particular flow sheet based on the vehicle of three-dimensional modeling;
Figure 6 shows that the automatic identification processing system structural representation of the traffic hazard that the present invention is based on video.
Fig. 7 is the synoptic diagram under the RGB coordinate system.
Embodiment
In order to make purpose of the present invention, technical scheme and advantage clearer, reach embodiment in conjunction with the following drawings, the traffic hazard that the present invention is based on video is identified disposal route and system is further elaborated automatically.Concrete enforcement described herein is only in order to explaining the present invention, and is not used in restriction the present invention.
Disposal route and system are identified in traffic hazard based on video of the present invention automatically, by the analysis to the road video monitoring image, realize the automatic identification of road traffic accident, and in time traffic accident information is sent to traffic control center, thereby reduce the loss that traffic hazard brings.
Describe the traffic hazard based on video of the present invention below in detail and identify disposal route automatically, as shown in Figure 1, comprise different automobile types is set up its three-dimensional feature storehouse respectively, obtain the road sequence of video images, prospect vehicle based on background modeling separates, based on the vehicle outline recognizer of three-dimensional modeling, based on the improvement SIFT feature recognition algorithms of the vehicle of three-dimensional modeling.Its treatment step is specific as follows:
S11 obtains the Real-time Road sequence of video images by video camera;
S12 based on background modeling, utilizes the background subtraction point-score that prospect vehicle and the road background that camera photographs is separated, and extracts the prospect vehicle;
S13, employing is based on the method for image difference, namely at first utilize 8 neighborhood communicating methods that moving target is carried out mark respectively, then the moving target that marks is calculated the size of its barycenter and connected region respectively, last and former frame image compares, judged whether that target is offset, and identifies stationary vehicle with this;
S14, use the canny operator extraction to go out the stationary vehicle profile that identifies among the step S13, obtain angle point in the contour curve as unique point, with the vehicle ' s contour Feature Points Matching in the three-D profile feature database, at first identify vehicle, according to profile whether deformation takes place again and judge whether to get into an accident, as judge to get into an accident and then report to the police, otherwise enter step S15, judge by improving the SIFT unique point;
S15 extracts the improvement SIFT unique point of vehicle, mates with local feature point in the vehicle three-dimensional feature storehouse, and comprehensively judges whether to get into an accident according to above-mentioned matching result.As traffic hazard takes place, and system is sent to district, city commander Surveillance center with traffic accident information and video by grade rapidly, and reports to the police, and makes that traffic accident can obtain rapidly handling, and personnel obtain relief.
As shown in Figure 2, the detailed process that different automobile types is set up its three-dimensional feature storehouse respectively may further comprise the steps:
At first to gather the three-dimensional image sequence for the treatment of the modeling vehicle, set up vehicle three-dimensional feature storehouse.This is the traffic accident that takes place owing to each, and the attitude of vehicle is immesurable, thereby need set up vehicle three-dimensional feature storehouse.In space coordinates, vehicle is positioned at coordinate origin, at first vehicle remains unchanged in Y-axis and Z axle formation plane, along on the plane that X-axis and Y-axis constitute clockwise or be rotated counterclockwise certain angle θ, gather a photograph frame of vehicle this moment, and will gather two features (the little vehicle ' s contour feature of this angle and improve the SIFT feature) set back warehouse-in of vehicle on this photograph frame, and then gather a photograph frame again behind the rotation θ, gather two characteristic sets and the warehouse-in of vehicle on this photograph frame.Rotate a circle like this and set up after the characteristic of correspondence, constitute on the plane in Y-axis and Z axle, around X-axis, behind the anglec of rotation ε, repeating above said collection one photograph frame extracts feature and warehouse-in until 360 ° of ε rotations.In this way, then can obtain the colourful attitude image of multi-angle of vehicle.Need explanation, anglec of rotation θ and ε can not be too big, and to prevent the missing Partial Feature point in the rotary course, the θ of this patent, the ε anglec of rotation are for adopting 11 °.
After obtaining the multi-angle image of vehicle, extract vehicle ' s contour, obtain the angle point of contour curve as the unique point of vehicle ' s contour.Obtain the improvement SIFT feature of vehicle in the contour curve simultaneously.The unique point of obtaining is described, is output as proper vector most.When obtaining feature, need add the angle information of feature in its feature, with the combination of the characteristic information after convenient.
Finish after the feature extraction, combination is screened and put in order to the feature of the image that all angles were photographed, and just can set up the three-dimensional feature model, and the proper vector after screening and the arrangement combination is deposited in the database.
As shown in Figure 3, the prospect vehicle separation detailed process based on background modeling comprises step:
At first, for colour imagery shot, adopt the background modeling method based on color space cluster road model.
Theoretical according to the color distortion, the pixel of video present frame is carried out cluster one by one, establish c
k=(R
k, G
k, B
k)
TBe the coded word of a pixel, I
kBe c
kBrightness, define the interval right cylinder that for this point and initial point line is spool of its distortion, wherein, Δ
cBe color distortion radius;
To the two continuous frames image f that collects
t, f
T+1Carry out difference, f
T+1?f
tObtain invariant region v wherein
T+1, each pixel in the invariant region is arranged following parameters, cluster centre c, brightness distortion radius Δ
IWith color distortion radius Δ
c, the subclass weights omega, maximum subclass number M gets sequence of video images first frame as initial back-ground model, with the color vector v of each location of pixels in this frame as its first cluster centre c
1, such weights omega is set simultaneously
1=1.Calculate the current pixel vector sum and had the distortion difference D of all cluster centres, and choose minimum value D wherein
MinAnd corresponding subclass k;
Be illustrated in fig. 7 shown below a certain pixel x of current video image
i=(R, G, B)
TTo c
kWith the distance, delta C of the line of rgb space true origin less than color distortion radius Δ
C, can think that this pixel satisfies color distortion rule, Δ C is the color distortion value of this pixel;
The brightness distortion rule is the brightness value I with a certain pixel of current video image
iWith coded word c
k=(R
k, G
k, B
k)
TBrightness I
kThe absolute value delta I=|I of difference
i-I
k| as the brightness distortion value, when Δ I less than brightness distortion radius Δ
IThe time, then this pixel satisfies such brightness distortion rule;
If D
MinSatisfy the clustering criteria of being formed jointly by color distortion rule and brightness distortion rule, show that current pixel belongs to subclass k, upgrades according to following formula the parameter of subclass so;
c
k,t+1(x,y)=(1-α)c
k,t(x,y)+αv
t+1(x,y)
In the formula: c
K, t+1(x, y)---(x y) locates k the cluster centre after the subclass renewal to pixel; c
K, t(x, y)---(x y) locates k subclass and upgrades preceding cluster centre pixel; α---learning rate, value are 0.03;
ω
k,t+1=(1-α)ω
k,t+α
Wherein, ω
K, t+1---the weight after k subclass upgraded; ω
K, t---the weight before k subclass upgraded;
If D
MinDo not satisfy clustering criteria, show that current pixel does not belong to any one already present subclass, if current subclass number is then added new subclass less than the maximum subclass number of setting, cluster centre is made as the current pixel proper vector, weight initialization ω
0=0.2; Otherwise, replace with the current pixel proper vector and to have had in the cluster centre minimum one of weight, also its weight is initialized as ω
0=0.2.
ω
K, t+1=(1-ω
0) ω
K, t, to other subclass weight decay.
To each location of pixels, sort to already present subclass is descending according to weights omega, and according to
Select the reasonable description of model as a setting of qualified top n subclass.If namely current pixel belongs to the top n subclass, can think that then it is the road background pixel, otherwise be foreground pixel.
For the black and white camera, adopt the background modeling method based on mixed Gauss model.
The basic thought of mixed Gaussian background modeling is: all use a plurality of Gauss models to represent the state that this pixel changes along with the variation of time simultaneously to the residing position of each pixel in video.In the mixed Gaussian background modeling, can select for use 3 to 5 single Gauss models to describe the feature of a certain pixel jointly usually.Each all uses a Gaussian function to represent in this K state, and according to the storage of sorting from big to small of its possibility that becomes background.Mixed Gauss model is described below:
Wherein, f
t--the t two field picture; ω
I, t---t frame f
2The weight of individual Gaussian distribution, and
η (f
t, μ
I, t,
)---i gauss of distribution function of t frame.
The renewal of mixed Gauss model mainly is that Gauss's parameter of describing its distribution is upgraded, in renewal process, need to consider simultaneously parameter and the parameter weighted value in model, whole process is more complicated also, also need sort again according to the size of weighted value at last.In shooting process, video is brought in constant renewal in, a new two field picture can constantly join in the video, this moment, mixed Gauss model was also in constantly upgrading, its basic thought is that the pixel following formula in the image that will newly obtain compares, if meet the distribution of mixed Gauss model then think that this point is background dot, if instead do not meet model, then regard as the impact point of motion.
According to matching result, each parameter of mixed Gauss model is upgraded, comprise weight, expectation and variance.ω
I, t+1=(1-α
1) ω
I, t+ α
1M
I, t+1, μ
I, t+1=(1-ρ) μ
I, t+ ρ * T
T+1,
α wherein
1(0≤α
1≤ 1) is the background learning rate, α
1Determining the speed of background model renewal speed, α
1Greatly then renewal speed is fast, α
1Little then renewal speed is slow.M
I, t+1Being illustrated in the matching degree of t+1 moment pixel color value and i Gauss model, is 1 during coupling, and not matching then is 0, and for unmatched Gauss model, its expectation and variance all remain unchanged, and the model for coupling upgrades it.ρ=α
1η (x
t| u
h, σ
k) be the Gauss model study factor, the speed of expression Gauss model parameter renewal.If do not have Gauss model and current picture value to be complementary, the Gauss model of weights minimum will be substituted so, and the expectation of new model is the current pixel color value, and variance is preset as a bigger initial value, and weight is preset as less initial value.
As shown in Figure 4, the vehicle outline recognizer concrete steps based on three-dimensional modeling comprise:
Use the canny operator to carry out edge extracting at first compute gradient value and deflection.Ask for the gradient M in x direction and y direction of foreground image respectively
xAnd M
y
Asking for gradient can carry out convolution in image by 3 * 3 templates and finish:
Grad is:
Gradient direction angle: θ=arctan (M
y/ M
x)
Be four direction with 0 °～360 ° gradient direction angle merger;
Non-maximization inhibition and hysteresis thresholdization obtain vehicle ' s contour curve f (x).
Contour curve f (x) is got its angle point, as the unique point of contour curve.
With the contour curve Feature Points Matching in the 3 d model library.
At first identify vehicle, criterion of identification is as follows:
The ratio of matching characteristic number and total all number of features is more than or equal to κ (κ value of the present invention is 0.6).
Judge again whether contour curve deformation takes place with the unique point of real-time contour curve f (x), unique point point coupling with contour curve f (x) ' in the 3 d model library, the ratio of matching characteristic number and all number of features is less than λ (the λ value is 0.9), then be judged to be traffic accident, and report to the police, otherwise enter next step, judge by improving the SIFT unique point.
As shown in Figure 5, the improvement SIFT feature recognition algorithms concrete steps based on the vehicle of three-dimensional modeling comprise:
Extract vehicle characteristics, be specially: at first use difference of Gaussian operator (DoG) to describe the multiple dimensioned expression of image, just generate the DoG metric space of image.
Shown in DoG is defined as follows: G (x, y, k σ)-G (x, y, σ)
Wherein: Gaussian function G (x, y, σ) be expressed as follows shown in:
For each point, determine whether it is extreme point.By reaching diagonal line eight points altogether up and down with it by point, also have 18 abutment points of levels to compare to determine whether it is extreme point.If determining this point is extreme point, this point is exactly unique point so, and can be according to this principal direction of the gradient calculation of its neighborhood;
Unique point is described, in improving SIFT unique point describing method, get around the unique point size and be 8 * 8 neighborhood, form 8 concentric circless, the accumulated value of the gradient weighting mould value of its 8 directions of statistics in each concentric circles scope, 8 dimensional vectors of choosing annular region from the inside to the outside in order respectively form final proper vector.Therefore, in improving the SIFT algorithm, use the vector of 64 dimensions to describe unique point.
Mate with the feature in the model bank, extracted the coupling that characteristics of image will carry out feature afterwards, the characteristics of image that extracts is with the formal description of vector, as a P
1(x
11, x
12..., x
1n) and P
2(x
21, x
22..., x
2n), the feature identification of this moment has uniqueness, identification point and be identified a little between have the highest similarity.
Point P
1(x
11, x
12..., x
1n) and P
2(x
21, x
22..., x
2n) between Euclidean distance be expressed as follows shown in:
Whether statistics obtains minimum distance and thinks that the similarity between two proper vectors that produce this minor increment is the highest then, after the coupling of carrying out feature, get into an accident by the ratio in judgement vehicle of matching characteristic number with total all number of features.
System of the present invention comprises communication subsystem, information storage subsystem, accident recognition subsystem and Incident Management subsystem as shown in Figure 6;
Wherein, described communication subsystem, 1) be used for the video sequence of camera acquisition is transferred to workstation; 2) be used for containing the accident transmission of video images to the server of district commander Surveillance center haveing through the workstation processing and identification; 3) finish communication between district, city commander Surveillance center and the server;
Information storage subsystem, the original video of storage traffic accident period of right time, the identifying information that corresponding video reflects the traffic accident grade, traffic accident be dot information and relevant rehabilitation information relatively;
The accident recognition subsystem, 1) the identification traffic accident takes place, and intercepting comprises the video information of traffic accident generating process; 2) identify intensity grade according to relevant specification or regulation;
The Incident Management subsystem is used for, and 1) the traffic accident warning; 2) traffic accident ranking compositor; 3) accident rehabilitation; 4) historical information inquiry, statistics.
Above-mentioned accident recognition subsystem comprises: background modeling module, prospect vehicle extraction module, vehicle ' s contour identification module and vehicle improve SIFT feature identification module;
Background modeling module: utilize the road video image information, adopt respectively based on the method for color space cluster with based on gauss hybrid models for colored and black and white camera and carry out background modeling;
Prospect vehicle extraction module: utilize the background difference algorithm, present frame and background are done difference, obtain the prospect vehicle, utilize eight connected domain algorithms that moving target is separately identified;
Vehicle ' s contour identification module: use the canny operator to obtain the profile of prospect vehicle, and mate with the three-dimensional vehicle ' s contour storehouse of prior foundation, at first identify vehicle, then if vehicle ' s contour deformation is obvious, be judged to be traffic accident, otherwise use local SFIT feature to differentiate;
Vehicle improves SIFT feature identification module: extract vehicle and improve the SIFT feature, mate with the feature database of prior foundation, judge whether to be traffic accident.
System of the present invention, can realize the responsibility of the transmission of the identification of traffic hazard and processing, accident information and preservation, traffic accident video playback, auxiliary identification accident etc., for traffic administration provides convenient, effective help, and the further investigation of intelligent transportation system is contributed to some extent.
More than specific embodiments of the invention are described and illustrate these embodiment only are exemplary, and be not used in and limit the invention, the present invention should make an explanation according to appended claim.
Claims (9)
1. identify disposal route automatically based on the traffic hazard of video, it is characterized in that, specifically may further comprise the steps:
Step S10: different automobile types is set up its three-dimensional feature storehouse respectively;
Step S11: obtain the road sequence of video images;
Step S12: the prospect vehicle based on background modeling separates;
Step S13: the static target of based target barycenter displacement is judged;
Step S14: based on the vehicle outline recognizer of three-dimensional modeling, identify vehicle, according to profile whether deformation takes place again and judge whether to get into an accident, as judge to get into an accident and then report to the police, otherwise enter step S15, judge by improving the SIFT unique point;
Step S15: based on the improvement SIFT feature recognition algorithms of the vehicle of three-dimensional modeling: extract the improvement SIFT unique point of vehicle, mate with local feature point in the vehicle three-dimensional feature storehouse, and comprehensively judge whether to get into an accident according to matching result.
2. traffic hazard automatic identifying method according to claim 1 is characterized in that, among the described step S10 different automobile types is set up its three-dimensional feature storehouse step respectively and comprises:
S101) obtain different automobile types at the image of the different attitude angles of three dimensions;
S102) extract the unique point of vehicle outline in the image;
S103) extract the improvement SIFT unique point in the vehicle closed curve that outline surrounds in the image;
S104) use the above-mentioned two category features point that extracts to set up vehicle three-dimensional feature storehouse.
3. traffic hazard automatic identifying method according to claim 1 is characterized in that, obtaining the road sequence of video images among the described step S11 is to obtain the road video image by setting up single fixed cameras with a certain interval at road.
4. traffic hazard automatic identifying method according to claim 1 is characterized in that, prospect vehicle separating step comprises among the described step S12:
S121) for colour imagery shot, adopt the background modeling method based on color space cluster road model; For the black and white camera, adopt the background modeling method based on mixed Gauss model;
S122) obtain prospect and background image based on the image difference point-score;
S123) utilizing morphology to step S122) the vehicle segmentation result improves.
5. traffic hazard automatic identifying method according to claim 1 is characterized in that, the stationary vehicle identification step based on image difference among the described step S13 comprises:
S131) utilize 8 neighborhood communicating methods that moving target is carried out mark respectively;
S132) moving target that marks is calculated the size of its barycenter and connected region respectively;
S133) compare with the former frame image, judged whether that target is offset.
6. traffic hazard automatic identifying method according to claim 1 is characterized in that, the vehicle outline recognizer step based on three-dimensional modeling among the described step S14 comprises:
S141) use the canny operator to obtain the outline of prospect vehicle;
S142) outline that obtains is carried out mathematical description, obtain its mathematic(al) representation;
S143) angle point in the contouring curve mates with the three-dimensional vehicle ' s contour storehouse of prior foundation as unique point;
S144) identify vehicle model, and judge whether the vehicle outline changes;
S145) change as outline, carry out traffic accident and report to the police, judge otherwise improve the SIFT feature by step S15.
7. traffic hazard automatic identifying method according to claim 1 is characterized in that, the vehicle improvement SIFT feature recognition algorithms step based on three-dimensional modeling among the described step S15 comprises:
S151) obtain the improvement SIFT feature of vehicle;
S152) mate with the three-dimensional feature storehouse of setting up in advance.
8. the automatic identification processing system of the traffic hazard based on video is characterized in that, this system comprises communication subsystem, information storage subsystem, accident recognition subsystem and Incident Management subsystem;
Wherein, described communication subsystem, 1) be used for the video sequence of camera acquisition is transferred to workstation; 2) be used for containing the accident transmission of video images to the server of district commander Surveillance center haveing through the workstation processing and identification; 3) finish communication between district, city commander Surveillance center and the server;
Described information storage subsystem, the original video of storage traffic accident period of right time, the identifying information that corresponding video reflects the traffic accident grade, traffic accident be dot information and relevant rehabilitation information relatively;
Described accident recognition subsystem, 1) the identification traffic accident takes place, and intercepting comprises the video information of traffic accident generating process; 2) identify intensity grade according to relevant specification or regulation;
Described Incident Management subsystem is used for 1) the traffic accident warning; 2) traffic accident ranking compositor; 3) accident rehabilitation; 4) historical information inquiry, statistics.
9. the automatic identification processing system of traffic hazard according to claim 8 is characterized in that, described accident recognition subsystem comprises: background modeling module, prospect vehicle extraction module, vehicle ' s contour identification module and vehicle improve SIFT feature identification module;
Background modeling module: utilize the road video image information, adopt respectively based on the method for color space cluster with based on gauss hybrid models for colored and black and white camera and carry out background modeling;
Prospect vehicle extraction module: utilize the background difference algorithm, present frame and background are done difference, obtain the prospect vehicle, utilize eight connected domain algorithms that moving target is separately identified;
Vehicle ' s contour identification module: use the canny operator to obtain the profile of prospect vehicle, and mate with the three-dimensional vehicle ' s contour storehouse of prior foundation, at first identify vehicle, then if vehicle ' s contour deformation is obvious, be judged to be traffic accident, otherwise use local SFIT feature to differentiate;
Vehicle improves SIFT feature identification module: extract vehicle and improve the SIFT feature, mate with the feature database of prior foundation, judge whether to be traffic accident.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310139545.6A CN103258432B (en) | 2013-04-19 | 2013-04-19 | Traffic accident automatic identification processing method and system based on videos |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310139545.6A CN103258432B (en) | 2013-04-19 | 2013-04-19 | Traffic accident automatic identification processing method and system based on videos |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103258432A true CN103258432A (en) | 2013-08-21 |
| CN103258432B CN103258432B (en) | 2015-05-27 |
Family
ID=48962312
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310139545.6A Expired - Fee Related CN103258432B (en) | 2013-04-19 | 2013-04-19 | Traffic accident automatic identification processing method and system based on videos |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103258432B (en) |
Cited By (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103473547A (en) * | 2013-09-23 | 2013-12-25 | 百年金海科技有限公司 | Vehicle target recognizing algorithm used for intelligent traffic detecting system |
| CN103793922A (en) * | 2013-09-12 | 2014-05-14 | 电子科技大学 | Real-time detection method for specific attitude |
| WO2016138640A1 (en) * | 2015-03-04 | 2016-09-09 | GM Global Technology Operations LLC | Systems and methods for assigning responsibility during traffic incidents |
| CN106373332A (en) * | 2016-09-30 | 2017-02-01 | 北京奇虎科技有限公司 | Vehicle-mounted intelligent alarm method and device |
| CN106548624A (en) * | 2016-11-03 | 2017-03-29 | 武汉理工大学 | A kind of car accident grade automatic identification and monitoring warning device, system and method |
| CN106781436A (en) * | 2016-12-19 | 2017-05-31 | 东软集团股份有限公司 | Traffic accident treatment method and device |
| CN107909113A (en) * | 2017-11-29 | 2018-04-13 | 北京小米移动软件有限公司 | Traffic-accident image processing method, device and storage medium |
| CN107933471A (en) * | 2017-12-04 | 2018-04-20 | 惠州市德赛西威汽车电子股份有限公司 | The method and vehicle-mounted automatic help system of accident active call rescue |
| CN108320517A (en) * | 2017-12-28 | 2018-07-24 | 浙江中新长清信息科技有限公司 | Car plate and vehicle identification system and monitoring server |
| CN108458691A (en) * | 2018-02-02 | 2018-08-28 | 新华智云科技有限公司 | A kind of collision checking method and equipment |
| CN108537125A (en) * | 2018-03-13 | 2018-09-14 | 周慧 | Automobile search system based on amendment type image procossing |
| CN108629963A (en) * | 2017-03-24 | 2018-10-09 | 纵目科技（上海）股份有限公司 | Traffic accident report method based on convolutional neural networks and system, car-mounted terminal |
| CN108921028A (en) * | 2018-06-01 | 2018-11-30 | 上海博泰悦臻电子设备制造有限公司 | The scene of a traffic accident regards acquisition method and system |
| CN108986468A (en) * | 2018-08-01 | 2018-12-11 | 平安科技（深圳）有限公司 | Processing method, device, computer equipment and the computer storage medium of traffic accident |
| CN108986474A (en) * | 2018-08-01 | 2018-12-11 | 平安科技（深圳）有限公司 | Fix duty method, apparatus, computer equipment and the computer storage medium of traffic accident |
| CN109118516A (en) * | 2018-07-13 | 2019-01-01 | 高新兴科技集团股份有限公司 | A kind of target is from moving to static tracking and device |
| CN109657573A (en) * | 2018-12-04 | 2019-04-19 | 联想(北京)有限公司 | Image-recognizing method and device and electronic equipment |
| CN109993976A (en) * | 2017-12-29 | 2019-07-09 | 技嘉科技股份有限公司 | Traffic accident monitors system and method |
| CN110782653A (en) * | 2018-11-07 | 2020-02-11 | 北京嘀嘀无限科技发展有限公司 | Road information acquisition method and system |
| WO2020061766A1 (en) * | 2018-09-25 | 2020-04-02 | 西门子股份公司 | Vehicle event detection apparatus and method, and computer program product and computer-readable medium |
| CN111199643A (en) * | 2018-11-20 | 2020-05-26 | 远创智慧股份有限公司 | Road condition monitoring method and system |
| CN111369807A (en) * | 2020-03-24 | 2020-07-03 | 北京百度网讯科技有限公司 | Traffic accident detection method, device, equipment and medium |
| CN111564058A (en) * | 2020-05-19 | 2020-08-21 | 杭州自动桌信息技术有限公司 | Device for monitoring parking lot condition and parking lot management system and management method |
| CN113378803A (en) * | 2021-08-12 | 2021-09-10 | 深圳市城市交通规划设计研究中心股份有限公司 | Road traffic accident detection method, device, computer and storage medium |
| CN114387785A (en) * | 2022-01-24 | 2022-04-22 | 陕西交通职业技术学院 | Safety management and control method and system based on intelligent highway and storable medium |
| WO2022099608A1 (en) * | 2020-11-13 | 2022-05-19 | 金序能 | Method for acquiring attribute category of traffic accident on highway |
| CN114241781B (en) * | 2021-11-22 | 2022-08-05 | 慧之安信息技术股份有限公司 | Automatic alarm function system based on traffic accident recognition |
| CN115148018A (en) * | 2021-03-31 | 2022-10-04 | 海信集团控股股份有限公司 | Traffic incident detection apparatus and method |
| CN116758494A (en) * | 2023-08-23 | 2023-09-15 | 深圳市科灵通科技有限公司 | Intelligent monitoring method and system for vehicle-mounted video of internet-connected vehicle |
| EP4229547A4 (en) * | 2020-11-20 | 2024-04-24 | Zhejiang Dahua Technology Co | Systems and methods for detecting traffic accidents |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0376800A1 (en) * | 1988-12-21 | 1990-07-04 | Serge Besnard | Automatic site monitoring process and apparatus |
| CN102073851A (en) * | 2011-01-13 | 2011-05-25 | 北京科技大学 | Method and system for automatically identifying urban traffic accident |
-
2013
- 2013-04-19 CN CN201310139545.6A patent/CN103258432B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0376800A1 (en) * | 1988-12-21 | 1990-07-04 | Serge Besnard | Automatic site monitoring process and apparatus |
| CN102073851A (en) * | 2011-01-13 | 2011-05-25 | 北京科技大学 | Method and system for automatically identifying urban traffic accident |
Non-Patent Citations (3)
| Title |
|---|
| 华莉琴，许维，王拓: "采用改进的尺度不变特征转换及多视角模型对车型识别", 《西安交通大学学报》 * |
| 拜佩，李金屏: "一种基于视频的交通事故检测方法", 《济南大学学报（自然科学版）》 * |
| 杨建国,尹旭全,方丽: "基于自适应轮廓匹配的视频运动车辆检测和跟踪", 《西安交通大学学报》 * |
Cited By (35)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103793922A (en) * | 2013-09-12 | 2014-05-14 | 电子科技大学 | Real-time detection method for specific attitude |
| CN103793922B (en) * | 2013-09-12 | 2016-07-06 | 电子科技大学 | A kind of particular pose real-time detection method |
| CN103473547A (en) * | 2013-09-23 | 2013-12-25 | 百年金海科技有限公司 | Vehicle target recognizing algorithm used for intelligent traffic detecting system |
| WO2016138640A1 (en) * | 2015-03-04 | 2016-09-09 | GM Global Technology Operations LLC | Systems and methods for assigning responsibility during traffic incidents |
| US20180047283A1 (en) * | 2015-03-04 | 2018-02-15 | Wende Zhang | Systems and methods for assigning responsibility during traffic incidents |
| CN106373332A (en) * | 2016-09-30 | 2017-02-01 | 北京奇虎科技有限公司 | Vehicle-mounted intelligent alarm method and device |
| CN106548624A (en) * | 2016-11-03 | 2017-03-29 | 武汉理工大学 | A kind of car accident grade automatic identification and monitoring warning device, system and method |
| CN106781436A (en) * | 2016-12-19 | 2017-05-31 | 东软集团股份有限公司 | Traffic accident treatment method and device |
| CN106781436B (en) * | 2016-12-19 | 2020-03-06 | 东软集团股份有限公司 | Traffic accident handling method and device |
| CN108629963A (en) * | 2017-03-24 | 2018-10-09 | 纵目科技（上海）股份有限公司 | Traffic accident report method based on convolutional neural networks and system, car-mounted terminal |
| CN107909113A (en) * | 2017-11-29 | 2018-04-13 | 北京小米移动软件有限公司 | Traffic-accident image processing method, device and storage medium |
| CN107933471A (en) * | 2017-12-04 | 2018-04-20 | 惠州市德赛西威汽车电子股份有限公司 | The method and vehicle-mounted automatic help system of accident active call rescue |
| CN107933471B (en) * | 2017-12-04 | 2019-12-20 | 惠州市德赛西威汽车电子股份有限公司 | Accident active calling rescue method and vehicle-mounted automatic help-seeking system |
| CN108320517A (en) * | 2017-12-28 | 2018-07-24 | 浙江中新长清信息科技有限公司 | Car plate and vehicle identification system and monitoring server |
| CN109993976A (en) * | 2017-12-29 | 2019-07-09 | 技嘉科技股份有限公司 | Traffic accident monitors system and method |
| CN108458691A (en) * | 2018-02-02 | 2018-08-28 | 新华智云科技有限公司 | A kind of collision checking method and equipment |
| CN108537125A (en) * | 2018-03-13 | 2018-09-14 | 周慧 | Automobile search system based on amendment type image procossing |
| CN108921028A (en) * | 2018-06-01 | 2018-11-30 | 上海博泰悦臻电子设备制造有限公司 | The scene of a traffic accident regards acquisition method and system |
| CN109118516A (en) * | 2018-07-13 | 2019-01-01 | 高新兴科技集团股份有限公司 | A kind of target is from moving to static tracking and device |
| CN108986474A (en) * | 2018-08-01 | 2018-12-11 | 平安科技（深圳）有限公司 | Fix duty method, apparatus, computer equipment and the computer storage medium of traffic accident |
| CN108986468A (en) * | 2018-08-01 | 2018-12-11 | 平安科技（深圳）有限公司 | Processing method, device, computer equipment and the computer storage medium of traffic accident |
| WO2020061766A1 (en) * | 2018-09-25 | 2020-04-02 | 西门子股份公司 | Vehicle event detection apparatus and method, and computer program product and computer-readable medium |
| CN110782653A (en) * | 2018-11-07 | 2020-02-11 | 北京嘀嘀无限科技发展有限公司 | Road information acquisition method and system |
| CN111199643A (en) * | 2018-11-20 | 2020-05-26 | 远创智慧股份有限公司 | Road condition monitoring method and system |
| CN109657573A (en) * | 2018-12-04 | 2019-04-19 | 联想(北京)有限公司 | Image-recognizing method and device and electronic equipment |
| CN111369807A (en) * | 2020-03-24 | 2020-07-03 | 北京百度网讯科技有限公司 | Traffic accident detection method, device, equipment and medium |
| CN111564058A (en) * | 2020-05-19 | 2020-08-21 | 杭州自动桌信息技术有限公司 | Device for monitoring parking lot condition and parking lot management system and management method |
| WO2022099608A1 (en) * | 2020-11-13 | 2022-05-19 | 金序能 | Method for acquiring attribute category of traffic accident on highway |
| EP4229547A4 (en) * | 2020-11-20 | 2024-04-24 | Zhejiang Dahua Technology Co | Systems and methods for detecting traffic accidents |
| CN115148018A (en) * | 2021-03-31 | 2022-10-04 | 海信集团控股股份有限公司 | Traffic incident detection apparatus and method |
| CN113378803A (en) * | 2021-08-12 | 2021-09-10 | 深圳市城市交通规划设计研究中心股份有限公司 | Road traffic accident detection method, device, computer and storage medium |
| CN114241781B (en) * | 2021-11-22 | 2022-08-05 | 慧之安信息技术股份有限公司 | Automatic alarm function system based on traffic accident recognition |
| CN114387785A (en) * | 2022-01-24 | 2022-04-22 | 陕西交通职业技术学院 | Safety management and control method and system based on intelligent highway and storable medium |
| CN116758494A (en) * | 2023-08-23 | 2023-09-15 | 深圳市科灵通科技有限公司 | Intelligent monitoring method and system for vehicle-mounted video of internet-connected vehicle |
| CN116758494B (en) * | 2023-08-23 | 2023-12-22 | 深圳市科灵通科技有限公司 | Intelligent monitoring method and system for vehicle-mounted video of internet-connected vehicle |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103258432B (en) | 2015-05-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103258432B (en) | Traffic accident automatic identification processing method and system based on videos | |
| CN107610114B (en) | optical satellite remote sensing image cloud and snow fog detection method based on support vector machine | |
| CN102708356B (en) | Automatic license plate positioning and recognition method based on complex background | |
| CN101389004B (en) | Moving target classification method based on on-line study | |
| Mi et al. | Research on regional clustering and two-stage SVM method for container truck recognition | |
| CN101339601B (en) | License plate Chinese character recognition method based on SIFT algorithm | |
| CN105005989B (en) | A kind of vehicle target dividing method under weak contrast | |
| CN105184271A (en) | Automatic vehicle detection method based on deep learning | |
| CN104951775A (en) | Video technology based secure and smart recognition method for railway crossing protection zone | |
| CN104378582A (en) | Intelligent video analysis system and method based on PTZ video camera cruising | |
| CN103605953A (en) | Vehicle interest target detection method based on sliding window search | |
| CN103065138A (en) | Recognition method of license plate number of motor vehicle | |
| CN106529461A (en) | Vehicle model identifying algorithm based on integral characteristic channel and SVM training device | |
| CN106778633B (en) | Pedestrian identification method based on region segmentation | |
| CN106503748A (en) | A kind of based on S SIFT features and the vehicle targets of SVM training aids | |
| CN104933398A (en) | vehicle identification system and method | |
| CN108520528B (en) | Mobile vehicle tracking method based on improved difference threshold and displacement matching model | |
| Mammeri et al. | North-American speed limit sign detection and recognition for smart cars | |
| CN104391966A (en) | Typical car logo searching method based on deep learning | |
| CN109977862B (en) | Recognition method of parking space limiter | |
| Zhang et al. | An on-road vehicle detection method for high-resolution aerial images based on local and global structure learning | |
| CN101996315B (en) | System and method for camera-based object analysis | |
| CN104517127A (en) | Self-learning pedestrian counting method and apparatus based on Bag-of-features model | |
| Balci et al. | Front-View Vehicle Damage Detection using Roadway Surveillance Camera Images. | |
| CN105574490A (en) | Vehicle brand identification method and system based on headlight image characteristics |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20150527 Termination date: 20190419 |