CN101923778A - Detection method of highway traffic congestion state based on video - Google Patents
Detection method of highway traffic congestion state based on video Download PDFInfo
- Publication number
- CN101923778A CN101923778A CN2009103068823A CN200910306882A CN101923778A CN 101923778 A CN101923778 A CN 101923778A CN 2009103068823 A CN2009103068823 A CN 2009103068823A CN 200910306882 A CN200910306882 A CN 200910306882A CN 101923778 A CN101923778 A CN 101923778A
- Authority
- CN
- China
- Prior art keywords
- light stream
- traffic
- traffic behavior
- optical flow
- state
- 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.)
- Pending
Links
Images
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention relates to a detection method of a traffic congestion state based on a video, and the traffic congestion detection usually focuses on the whole situation of traffic flows rather than traffic flow parameters of an individual vehicle. The directivity and regularity of movement of vehicles on a highway are obvious, an optical flow field is orderly and regular, and an optical flow vector reflects the size and direction of the traffic speed directly. The detection of the road state can be realized by studying variation law of intensity of road movement information of the optical flow in a video monitoring range. The invention provides the novel detection method of highway traffic congestion under monitoring of a macro wide-angle video, which comprises the following two steps: the first step: calculating the optical flow field of an overall image and eliminating optical flow points that do not satisfy a main direction; and then calculating an average value of the optical speeds after being screened, which is taken as a macro optical flow speed value of the vehicles; and the second step: judging whether traffic congestion occurs according to the average speed of the optical flow in combination with duration features of the traffic state.
Description
Technical field
The invention belongs to digital image processing field, be specifically related to a kind of highway traffic congestion state decision method based on video.
Background technology
The traffic congestion state means of identification has artificial cognition and discerns two kinds automatically, artificial cognition is mainly by artificial observations such as CCTV supervision and Smokey's reports, automatically identification mainly is to discern automatically by the traffic congestion plug that (Automatic TrafficCongestion Identification, ACI) algorithm is realized the automatic identification to the traffic congestion plug.Because artificial cognition can take many human resources, and is subjected to factor affecting such as weather and time, therefore, the ACI algorithm of the numerous and confused development of new in various countries is used for the identification of traffic congestion plug.
In the past, the researchist has developed many ACI algorithms, and wherein more famous have California algorithm and a McMaster algorithm.The California serial algorithm is that 1880s is developed by California, USA Department of Transportation, and this algorithm has been used for many years in adding the highway traffic congestion recognition system of downstate, and effect is relatively good.The McMaster algorithm is based on the catastrophe theory exploitation of traffic flow by department of civil engineering of Canadian McMaster university, this algorithm has utilized the measured data of Canadian Elizabethan street and the Burlington Skyway of Ontario to test, the result of test is extremely successful, afterwards this algorithm has been carried out on-line testing and further improvement.
Except these two kinds of algorithms of above introduction, also there are many researchists to be engaged in the research work of ACI algorithm aspect.Nineteen eighty-two Ahmad and Cook have proposed a kind of traffic congestion based on occupation rate Time series analysis method recognizer of blocking up, in this algorithm, utilize time series models that a variation tendency that detects occupation rate in several time intervals in website past is carried out match, and go out the scope of next time interval occupation rate by this model prediction, if it is too big that measured data tangible deviation occurs or departs from the predicted mean vote degree, this algorithm will be differentiated the generation traffic congestion and block up.Texas, USA transportation association utilized method for mode matching to develop a kind of algorithm that traffic congestion is blocked up under the low discharge state that is used to discern in 1979.This algorithm discipline registration of vehicle is through the time and the speed at upstream detection station, and can provide the time value estimation range of this vehicle through the adjacent downstream measuring station, utilizing the speed of upstream vehicle and the distance calculation between adjacent two measuring stations to go out then can be from the flow value of detected downstream station process in this time range, and the detected actual flow value of this flow value and detected downstream station contrasted, if difference greater than a given threshold value, is then judged the congested generation of blocking up is arranged on the highway section.The transportation of Britain and roadway experiment chamber (TRRL) have developed two kinds of algorithms in 1979.First kind of algorithm is used under the high occupation rate situation, is 100% when the occupation rate of a detecting device continues two seconds, and judging so has the congested generation of blocking up.Second kind of algorithm utilizes mode-matching technique, find out distance and be approximately driving vehicle between two measuring stations of 500 meters and estimate the interval average speed of vehicle to the time delay of expression patterns, if actual detected to the speed of a motor vehicle have one significantly to descend then judgement has the congested generation of blocking up.Willsky etc. and Cremer and Schutt utilize Kalman Filter Technology to estimate the parameter of express highway section traffic flow, when the super given threshold value of the variation of a certain parameter value, then judging has the congested generation of blocking up on this highway section, but this method also rests on laboratory stage.People such as Blosseville used image processing techniques to discern the generation that traffic congestion is blocked up on the highway in 1991, and in initial test, the method can identify 124 from 133 congested blocking up, and 9 mistakes only occurred and knew.This method is suitable for discerning the vehicle of stagnation of movement on curb and main carriageway, can also provide queue length simultaneously, and the method is also in further studying.Owing to exist inherent contact between discrimination and the misclassification rate, some researchists have begun statistical technique is applied in the automatic identification that traffic congestion blocks up, Levin in 1978 and Krause use adjacent two to detect between the websites occupation rate time difference as input feature vector, with " committee-machine " method that is similar to the neural network model structure congested blocking up discerned, but this method just is in laboratory stage.
At present, domesticly also be in the starting stage aspect the research of ACI algorithm, there are Research Institute of Highway, Ministry of Communications ITS center, Jilin University, Tongji University, Beijing Jiaotong University and Beijing University of Technology etc. in the unit that is engaged in this respect research work, but the ACI algorithm of comparative maturity is fewer, mainly is the continuation research to external algorithm.
The present invention by the optical flow computation technology, obtains macroscopical telecommunication flow information innovatively, carries out traffic congestion state in conjunction with the traffic behavior feature and detects, and overcome the limitation of the algorithm that detects, follows the tracks of based on single unit vehicle, and this kind method forefathers does not study.
Summary of the invention
Whether the objective of the invention is to judge traffic behavior, have congestion situation to take place thereby detect, the specific implementation process was divided into for two steps: one, utilize the optical flow field computing method, calculate the optical flow field vector of wagon flow, obtain macroscopical light stream average velocity; Adopt following steps to handle: 1. initial key feature point extraction, 2. at the macroscopical light stream speed of key feature point calculating.Two, judge that according to light stream velocity amplitude and traffic behavior protensive features traffic congestion state takes place; 1. judge it is to belong to any traffic behavior (smooth and easy, walk or drive slowly, block up) according to macroscopical light stream speed is preliminary, 2. finally determine traffic behavior according to the traffic behavior metastasis model, 3. set the time threshold that blocks up, the overtime threshold value is carried out jam alarming.Concrete technical scheme is as follows:
Calculate light stream average velocity
Because the vehicle movement rule of the road area under the same video monitoring generally has certain similarity, the direction of motion and the speed that comprise vehicle, and car speed to can be used as road condition (smooth and easy, crowded) the objective evaluation standard, therefore at all vehicles in the entire image, calculate the integral macroscopic light stream speed of all vehicles, just can roughly hold the integrality of traffic flow.Calculate traffic flow macroscopic view light stream speed method so the present invention proposes a kind of fusion Corner Detection and optical flow computation, this method mainly was divided into for two steps:
The first step is the initial key feature point extraction.Concrete grammar is: at first entire image is carried out optical flow computation and angle point and extracts, find out in the image be sharp movement point be again the pixel of strong angle point, as the initial key unique point.
Because it is insecure that the optical flow field at the uncontinuity of optical flow field and violation conservation assumed condition place distributes, if calculate its light stream at the pixel at non-vanishing place in the differential chart, the optical flow field that calculates whole moving object is more reliable.This is because they often corresponding to the bigger point of shade of gray, are set up and the optical flow field fundamental equation at these some places is approximate.After adopting this constraint measure, it is more reliable and accurate to make that the optical flow field that calculates distributes, and has also reduced calculated amount (this is because needn't calculate entire image, only needs the optical flow field at non-vanishing place among the calculated difference figure to distribute).Simultaneously, because optical flow computation is at global image, many light streams are unwanted in the optical flow field, particularly light stream is worth the noise spot of too small pixel, screening is also removed these unique points and is extracted our desired unique point, only handle, can improve computing velocity and accuracy rate widely at the bigger unique point of these light stream values.
Second step was the calculating of macroscopical light stream speed.Concrete grammar is: the optical flow field with Horn-Schunck algorithm computation general image determines light stream principal direction then, and removes the light stream point that does not satisfy principal direction; Use the light stream speed average after Lucas-Kanade pyramid algorith is calculated screening at last, as the light stream velocity amplitude of vehicle macroscopic view.
Because the unique point that the first step is extracted might not all be in vehicle region, and direction is also not necessarily identical with vehicle heading.In order only to calculate the unique point of vehicle, this paper defines the notion of light stream principal direction and carries out the screening operation of light stream principal direction, filters out and satisfies the unique point that meets light stream principal direction, limits unique point and is in site of road.Adopt optical flow algorithm to calculate the light stream intensity of being left unique point at last, and the light stream velocity amplitude of its mean value as the traffic flow macroscopic view.
Obscure light stream and noise to calculating the influence of traffic flow macroscopic view light stream speed in order to reduce as much as possible, calculate sightseeing flow velocity degree more accurately, at first adopt Horn-Schunck algorithm computation optical flow field value, through screening of optical flow field value threshold value and the very few noise spot of removal light stream value, calculate the direction of residue character point and add up all directions, and calculate principal direction.Then, utilize principal direction to remove the influence of obscuring light stream, remaining point as the key feature point, is calculated the light stream velocity amplitude of key feature point.In order to make calculating more accurate, use Lucas-Kanade pyramid algorith, calculation process is as follows:
Generate the m+1 layer Guassian pyramid structure of image, wherein the 0th layer is original image, and in the standard light stream (Lucas-Knadae algorithm) of top m calculated characteristics point;
When calculating the i layer:
Obtain i+1 layer unique point (k, light stream u l)
I+1(k, l) and v
I+1(k, l);
With bilinear interpolation calculated characteristics point (k, compensation light stream u l)
i *(k, l), v
i *(k, l);
With u
i *(k, l), v
i *(k l) multiply by 2;
Use u
i *(k, l), v
i *(k, l) the pyramid diagram picture of compensation i layer present frame;
Between the pyramid diagram picture after the compensation, use the light stream u ' of standard optical flow computation unique point
i(k, l) and v '
i(k, l);
Finally calculate the optical flow field of i layer unique point:
Repeated for (2) step, until the optical flow field that calculates final the 0th layer of (original image) all unique point.
By above-mentioned optical flow computation, get access to the light stream velocity amplitude of each unique point level and vertical two directions, respectively the mean value of calculated characteristics point level and vertical two direction light stream velocity amplitudes to all screening unique points
With
The flow process of the macroscopical light stream speed of whole calculating as shown in the figure.
Description of drawings
Fig. 1 is traffic congestion testing process figure.
Fig. 2 is macroscopical light stream speed calculation process flow diagram.
Fig. 3 is the free-moving traffic constitutional diagram.The light stream vector of red arrow remarked pixel point wherein.
Fig. 4 is the slow-moving traffic constitutional diagram.The light stream vector of red arrow remarked pixel point wherein.
Fig. 5 is the traffic behavior figure that blocks up for the road surface, left side.The light stream vector of red arrow remarked pixel point wherein.
Fig. 6 is macroscopical light stream velocity analysis figure of traffic behavior.Light stream speed when wherein green curve is represented to pass unimpeded state; Light stream speed when blue curve is represented to go slowly state; Light stream speed when red curve is represented congestion status.
Fig. 7 is traffic behavior model structure figure.Comprise three kinds of traffic behaviors, be respectively the state of passing unimpeded, the state of going slowly, congestion status.
Fig. 8 is traffic behavior transformational relation figure.It is intended to expression, the conversion of traffic behavior has two kinds of possibilities, is respectively: state one, passes unimpeded--state of going slowly--congestion status; Two, congestion status--state of going slowly--passes unimpeded state.The state of passing unimpeded--congestion status can not appear, perhaps congestion status--and state passes unimpeded.
Fig. 9 is a table 1, the traffic behavior check matrix.From table, can get, can check the current detection state, obtain " detected state after checking ", thereby proofread and correct the mistake that logically occurs by " detected state last time " and " current detection state ".
Embodiment
Feature in conjunction with traffic behavior judges whether that traffic congestion has occurred
The generation of judging traffic congestion specifically comprises following two steps.
1. the macroscopical light stream velocity characteristic under each scene
Concrete grammar is: since through research learn pass unimpeded and go slowly between macroscopical light stream speed differ bigger, so at first calculate the light stream velocity amplitude of vehicle macroscopic view, and tentatively judge traffic behavior in the light stream threshold value of going slowly according to passing unimpeded of being studied before.
Suppose that at first the macroscopical light stream speed under the different traffic scene is different, three groups of following experiment confirms this hypothesis, the highway of source video sequence reality is tested in three groups of experiments respectively under two kinds of traffic conditions, video size is 320*240, and frame per second is 25fps.Road traffic state in one of them video is to pass unimpeded, as Fig. 3; Road traffic state in video is to go slowly, as Fig. 4; Road traffic state in last video is to go slowly, as Fig. 5, and three key feature points that figure Smalt is punctuated and represented said extracted, red arrow is partly represented corresponding vehicle light stream.Calculate the macroscopical light stream speed under above-mentioned three kinds of situations, by 9 groups of macroscopic view light stream speed that adjacent two frames in continuous 10 frames calculate, the gained result relatively sees Fig. 6.
From the result of Fig. 6, pass unimpeded and go slowly between macroscopical light stream speed differ bigger, result of study shows that when passing unimpeded state, its macroscopical light stream speed is generally greater than 1 pixel/frame; When going slowly state, its macroscopical light stream speed is generally greater than 0.25 pixel/frame, less than 1 pixel/frame; Macroscopical light stream speed average generally is less than 0.25 pixel/frame under the jam situation, is almost 0 under a lot of situations.
2. traffic behavior metastasis model research
The concrete grammar of final definite traffic behavior is: at first study three kinds of traffic behaviors and determine basic state model, and determine triangular relation promptly: occurring the first time of congestion status must be through passing unimpeded and the state of walking or drive slowly; To the state of passing unimpeded, the jogging state must be passed through in the centre from congestion status, has also embodied a kind of state of process.Secondly finally determine traffic behavior according to the traffic behavior metastasis model.
By above-mentioned macroscopical light stream velocity analysis, can distinguish by macroscopical light stream speed and to pass unimpeded, go slowly and jam situation.But, change in the optical flow field of image owing to calculate traffic flow macroscopic view light stream speed dependent, if optical flow field value is too small, cause that mainly due to two kinds of reasons a kind of is the situation that road does not have car, another kind is the situation of road vehicle obstruction.
Because the limitation of optical flow computation generally can obtain identical result in both cases, obviously, only rely on the principal direction of aforementioned calculation and two macroscopical parameters of macroscopical light stream speed to be difficult to this class situation of accurate description, in order to address this problem.Three kinds of traffic behaviors of the present invention's research are determined basic state model, and determine triangular relation, as shown in Figure 7.The supposition three kinds of states that pass unimpeded, go slowly, block up are mutually independently in the model, do not exist both to coexist and the situation of three's coexistence.
Have hysteresis quality because the traffic behavior of wagon flow shifts, thereby produce state conversion process Zhao transition state.Because the transition state duration is shorter relatively, and the determination flow when going slowly state is more similar, generally transition state is divided into the state of going slowly.According to transformational relation and process between three kinds of states, the method for traffic behavior branching decision is proposed, based on the following:
1) existence shifts between the traffic behaviors such as passing unimpeded, go slowly, block up: pass unimpeded and block up to going slowly, going slowly to, block up to go slowly and go slowly to and pass unimpeded etc., can't have that crossing from passing unimpeded blocks up or block up crosses the state of passing unimpeded.
2) different conditions is generally lasting different, and the whole duration of generally passing unimpeded is the longest.In addition, the process of traffic behavior conversion has the regular hour.
Three kinds of traffic behavior transformational relations as shown in Figure 8.
After determining three kinds of state exchange relations, as can be seen from Figure 8 go slowly be in pass unimpeded and congestion status between transition period, if cycle detection time is shorter, then can be according to last time traffic behavior and current traffic behavior, determine the correctness of the traffic behavior of current detection, it is as shown in table 1 to check the concrete relation table in back.
By to the researching and analysing of above-mentioned traffic congestion, can detect jam situation according to the macroscopical light stream speed calculated and the checking of transformational relation between traffic behavior.
The setting time threshold that blocks up, the concrete grammar that the overtime threshold value is carried out jam alarming is: when detecting traffic behavior and occur detecting jam situation, may be instantaneous generation owing to block up, be paroxysmal.Blocking up of short time do not need to report to the police, and the inventive method is set a duration at interval, just reports to the police after only detection is blocked up and continued for some time.
The inventive method mainly contains following characteristics:
(1) detailed analysis is carried out in light stream under the traffic scene, proposed the concept of light stream principal direction and the computational methods of macroscopical light stream speed.
(2) by the optical flow computation technology, obtain macro-traffic optical flow field information, realize detecting carrying out traffic congestion, overcome the limitation of the algorithm that detects, follows the tracks of based on single unit vehicle.
(3) propose the traffic behavior metastasis model, and followed according to transformational relation and process between three kinds of states, formulated the method for traffic behavior branching decision.
Claims (6)
1. detection method of highway traffic congestion state based on video is characterized in that this method may further comprise the steps:
A. utilize the optical flow field computing method, calculate the optical flow field vector of wagon flow, obtain macroscopical light stream average velocity; Adopt following steps to handle: 1. initial key feature point extraction, 2. at the macroscopical light stream speed of key feature point calculating;
B. judge that according to light stream velocity amplitude and traffic behavior protensive features traffic congestion state takes place; 1. judge it is to belong to any traffic behavior according to macroscopical light stream speed is preliminary, as smooth and easy, walk or drive slowly, block up, 2. finally determine traffic behavior according to the traffic behavior metastasis model, 3. set the time threshold that blocks up, the overtime threshold value is carried out jam alarming.
2. traffic behavior detection method according to claim 1, it is characterized in that among the described step a that 1. the concrete grammar of initial key feature point extraction is: at first entire image is carried out optical flow computation and angle point extracts, find out in the image be sharp movement point be again the pixel of strong angle point, as the initial key unique point.
3. traffic behavior detection method according to claim 1, it is characterized in that among the described step a? the concrete grammar that calculates macroscopical light stream speed at key feature point is: with the optical flow field of Horn-Schunck algorithm computation general image, determine light stream principal direction then, and remove the light stream point that does not satisfy principal direction; Use the light stream speed average after Lucas-Kanade pyramid algorith is calculated screening at last, as the light stream velocity amplitude of vehicle macroscopic view.
4. traffic behavior detection method according to claim 1, it is characterized in that 1. judging it is to belong to any traffic behavior among the described step b according to macroscopical light stream speed is preliminary, concrete grammar smooth and easy, that walk or drive slowly, block up is: since through research learn pass unimpeded and go slowly between macroscopical light stream speed differ bigger, so at first calculate the light stream velocity amplitude of vehicle macroscopic view, and tentatively judge traffic behavior in the light stream threshold value of going slowly according to passing unimpeded of being studied before.
5. traffic behavior detection method according to claim 1, it is characterized in that 2. determining finally that according to the traffic behavior metastasis model concrete grammar of traffic behavior is among the described step b: at first study three kinds of traffic behaviors and determine basic state model, and determine triangular relation promptly: occur the first time of congestion status passing through the state that passes unimpeded with jogging; To the state of passing unimpeded, the jogging state must be passed through in the centre from congestion status, has also embodied a kind of state of process, secondly finally determines traffic behavior according to the traffic behavior metastasis model.
6. traffic behavior detection method according to claim 5, it is characterized in that 3. setting among the described step b time threshold that blocks up, the concrete grammar that the overtime threshold value is carried out jam alarming is: when occurring detecting jam situation for the detection traffic behavior, owing to block up may be instantaneous generation, be paroxysmal, blocking up of short time do not need to report to the police, and the inventive method is set a duration at interval, just reports to the police after only detection is blocked up and continued for some time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009103068823A CN101923778A (en) | 2009-09-11 | 2009-09-11 | Detection method of highway traffic congestion state based on video |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009103068823A CN101923778A (en) | 2009-09-11 | 2009-09-11 | Detection method of highway traffic congestion state based on video |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101923778A true CN101923778A (en) | 2010-12-22 |
Family
ID=43338680
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009103068823A Pending CN101923778A (en) | 2009-09-11 | 2009-09-11 | Detection method of highway traffic congestion state based on video |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101923778A (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102176285A (en) * | 2011-02-28 | 2011-09-07 | 江苏怡和科技股份有限公司 | Method for judging behavior patterns of vehicles in video stream |
| CN102254428A (en) * | 2011-04-28 | 2011-11-23 | 崔志明 | Traffic jam detection method based on video processing |
| CN102938203A (en) * | 2012-11-06 | 2013-02-20 | 江苏大为科技股份有限公司 | Basic traffic flow parameter based automatic identification method for traffic congestion states |
| CN103021181A (en) * | 2012-12-30 | 2013-04-03 | 西安费斯达自动化工程有限公司 | Traffic congestion monitoring and predicting method based on macro discrete traffic flow model |
| CN103413325A (en) * | 2013-08-12 | 2013-11-27 | 大连理工大学 | Vehicle speed identification method based on vehicle body feature point positioning |
| CN104282165A (en) * | 2013-07-12 | 2015-01-14 | 深圳市赛格导航科技股份有限公司 | Early-warning method and device for road segment congestion |
| CN105574895A (en) * | 2016-01-05 | 2016-05-11 | 浙江博天科技有限公司 | Congestion detection method during the dynamic driving process of vehicle |
| CN105608431A (en) * | 2015-12-22 | 2016-05-25 | 杭州中威电子股份有限公司 | Vehicle number and traffic flow speed based highway congestion detection method |
| CN107301369A (en) * | 2017-09-04 | 2017-10-27 | 南京航空航天大学 | Road traffic congestion analysis method based on Aerial Images |
| CN108162858A (en) * | 2016-12-07 | 2018-06-15 | 杭州海康威视数字技术股份有限公司 | Vehicle-mounted monitoring apparatus and its method |
| CN108615358A (en) * | 2018-05-02 | 2018-10-02 | 安徽大学 | A kind of congestion in road detection method and device |
| CN109147331A (en) * | 2018-10-11 | 2019-01-04 | 青岛大学 | A kind of congestion in road condition detection method based on computer vision |
| CN109410598A (en) * | 2018-11-09 | 2019-03-01 | 浙江浩腾电子科技股份有限公司 | A kind of traffic intersection congestion detection method based on computer vision |
| CN109766867A (en) * | 2019-01-22 | 2019-05-17 | 长沙智能驾驶研究院有限公司 | Travel condition of vehicle determines method, apparatus, computer equipment and storage medium |
| CN110598511A (en) * | 2018-06-13 | 2019-12-20 | 杭州海康威视数字技术股份有限公司 | Method, device, electronic equipment and system for detecting green light running event |
| CN112434260A (en) * | 2020-10-21 | 2021-03-02 | 北京千方科技股份有限公司 | Road traffic state detection method and device, storage medium and terminal |
-
2009
- 2009-09-11 CN CN2009103068823A patent/CN101923778A/en active Pending
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102176285A (en) * | 2011-02-28 | 2011-09-07 | 江苏怡和科技股份有限公司 | Method for judging behavior patterns of vehicles in video stream |
| CN102254428A (en) * | 2011-04-28 | 2011-11-23 | 崔志明 | Traffic jam detection method based on video processing |
| CN102938203A (en) * | 2012-11-06 | 2013-02-20 | 江苏大为科技股份有限公司 | Basic traffic flow parameter based automatic identification method for traffic congestion states |
| CN103021181A (en) * | 2012-12-30 | 2013-04-03 | 西安费斯达自动化工程有限公司 | Traffic congestion monitoring and predicting method based on macro discrete traffic flow model |
| CN103021181B (en) * | 2012-12-30 | 2014-10-08 | 西安费斯达自动化工程有限公司 | Traffic congestion monitoring and predicting method based on macro discrete traffic flow model |
| CN104282165A (en) * | 2013-07-12 | 2015-01-14 | 深圳市赛格导航科技股份有限公司 | Early-warning method and device for road segment congestion |
| CN103413325A (en) * | 2013-08-12 | 2013-11-27 | 大连理工大学 | Vehicle speed identification method based on vehicle body feature point positioning |
| CN103413325B (en) * | 2013-08-12 | 2016-04-13 | 大连理工大学 | A kind of speed of a motor vehicle authentication method based on vehicle body positioning feature point |
| CN105608431A (en) * | 2015-12-22 | 2016-05-25 | 杭州中威电子股份有限公司 | Vehicle number and traffic flow speed based highway congestion detection method |
| CN105574895A (en) * | 2016-01-05 | 2016-05-11 | 浙江博天科技有限公司 | Congestion detection method during the dynamic driving process of vehicle |
| CN108162858A (en) * | 2016-12-07 | 2018-06-15 | 杭州海康威视数字技术股份有限公司 | Vehicle-mounted monitoring apparatus and its method |
| CN107301369A (en) * | 2017-09-04 | 2017-10-27 | 南京航空航天大学 | Road traffic congestion analysis method based on Aerial Images |
| CN108615358A (en) * | 2018-05-02 | 2018-10-02 | 安徽大学 | A kind of congestion in road detection method and device |
| CN110598511A (en) * | 2018-06-13 | 2019-12-20 | 杭州海康威视数字技术股份有限公司 | Method, device, electronic equipment and system for detecting green light running event |
| CN109147331A (en) * | 2018-10-11 | 2019-01-04 | 青岛大学 | A kind of congestion in road condition detection method based on computer vision |
| CN109147331B (en) * | 2018-10-11 | 2021-07-27 | 青岛大学 | Road congestion state detection method based on computer vision |
| CN109410598A (en) * | 2018-11-09 | 2019-03-01 | 浙江浩腾电子科技股份有限公司 | A kind of traffic intersection congestion detection method based on computer vision |
| CN109766867A (en) * | 2019-01-22 | 2019-05-17 | 长沙智能驾驶研究院有限公司 | Travel condition of vehicle determines method, apparatus, computer equipment and storage medium |
| CN112434260A (en) * | 2020-10-21 | 2021-03-02 | 北京千方科技股份有限公司 | Road traffic state detection method and device, storage medium and terminal |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101923778A (en) | Detection method of highway traffic congestion state based on video | |
| CN104732075B (en) | A kind of Urban Road Traffic Accidents risk real-time predicting method | |
| CN104464295B (en) | A kind of overhead Entrance ramp intelligence restricted driving method based on video | |
| Hiribarren et al. | Real time traffic states estimation on arterials based on trajectory data | |
| CN104751642B (en) | A kind of advanced road real-time predictor method of traffic flow operation risk | |
| CN104464307B (en) | A kind of tunnel traffic based on video blocks up event automatic detection method | |
| CN104392607A (en) | Method and device for identifying traffic conditions | |
| Aljamal et al. | Real-time estimation of vehicle counts on signalized intersection approaches using probe vehicle data | |
| Jeng et al. | A high-definition traffic performance monitoring system with the inductive loop detector signature technology | |
| CN104408915A (en) | Traffic state parameter estimation method and system | |
| CN106156890B (en) | Detection of passenger flow and Forecasting Methodology and its system in a kind of urban track traffic channel | |
| KR101123967B1 (en) | Traffic congestion prediction system, prediction method and recording medium thereof | |
| Asamer et al. | Estimation of road capacity and free flow speed for urban roads under adverse weather conditions | |
| CN104575049B (en) | A kind of elevated ramp intellectual inducing method and device based on array radar | |
| Hawas et al. | Optimized multistage fuzzy-based model for incident detection and management on urban streets | |
| Salim et al. | Estimation of average space headway under heterogeneous traffic conditions | |
| Ma et al. | Prediction of arterial travel time considering delay in vehicle re-identification | |
| Margreiter | Fast and reliable determination of the traffic state using Bluetooth detection on German freeways | |
| US20130289864A1 (en) | Identifying impact of a traffic incident on a road network | |
| Rehborn et al. | Common traffic congestion features studied in USA, UK, and Germany based on kerner's three-phase traffic theory | |
| Brunauer et al. | Deriving driver-centric travel information by mining delay patterns from single GPS trajectories | |
| Detzer et al. | Analysis of traffic safety for cyclists: The automatic detection of critical traffic situations for cyclists | |
| Xia et al. | Predicting freeway travel time under incident conditions | |
| Chen et al. | Automatic freeway bottleneck identification and visualization using image processing techniques | |
| CN101587644A (en) | Automatic detection method for traffic accident on urban expressway based on non-continuous sliding sequence |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20101222 |