CN109754599B - Crossing traffic state identification method based on space-time analysis - Google Patents
Crossing traffic state identification method based on space-time analysis Download PDFInfo
- Publication number
- CN109754599B CN109754599B CN201811554437.4A CN201811554437A CN109754599B CN 109754599 B CN109754599 B CN 109754599B CN 201811554437 A CN201811554437 A CN 201811554437A CN 109754599 B CN109754599 B CN 109754599B
- Authority
- CN
- China
- Prior art keywords
- flow
- time
- traffic
- lane
- stage
- 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
Images
Abstract
The invention provides a crossing traffic state identification method based on space-time analysis, which comprises the steps of data acquisition, preprocessing, feature vector selection and definition and classifier design. Firstly, a characteristic vector of traffic state recognition is obtained through a space-time analysis technology, namely, the space traffic flow data and the time real-time signal data are subjected to stage-level refined deep fusion analysis to obtain a value parameter which is more fit with the traffic state, then the obtained characteristic vector is subjected to learning training, and the traffic state is automatically recognized.
Description
Technical Field
The patent relates to the technical field of traffic state identification through computer software.
Background
Urban road traffic congestion becomes one of the main social problems which puzzles various cities in the world, not only brings much inconvenience to daily work and life of people, but also restricts the increase of economy, accelerates the deterioration of urban environment and the like, and seriously influences the sustainable development of cities. The method and the device can timely and accurately identify the traffic jam in the road network, and have important significance for formulating a reasonable and effective traffic signal control dredging strategy. Finding out the characteristics which can effectively realize classification and identification and better fit the traffic state from a plurality of characteristics is a key first problem to be solved urgently.
Most of feature vectors selected by traditional traffic state identification are based on basic data of single-section flow and occupancy characteristics acquired by coil detection, geomagnetic detection and microwave detection or vehicle queuing characteristics in a single detection area, the data cannot finely depict vehicle dynamic information in a road junction canalization section, and the real-time traffic running state of the road junction cannot be represented well fundamentally, and most importantly, the data are isolated from signal release state data, and the result requirement (namely, the traffic signal control dredging strategy) of the traffic state identification is formulated, so that the identification result cannot directly guide the specific implementation of a signal control dredging strategy of related technicians.
Disclosure of Invention
The purpose of the invention is as follows: the invention provides a traffic signal coordination control method capable of automatically adjusting a period and a split green ratio, which is used for realizing trunk line coordination control with signal period duration and phase duration automatically adjusted under the condition that traffic flow meeting a coordination direction passes through an intersection once. The utilization rate of the green light at the intersection is fully improved, and the passing efficiency of the intersection is improved.
The technical scheme is as follows: in order to achieve the purpose, the invention can adopt the following technical scheme:
a crossing traffic state identification method based on space-time analysis comprises the following steps:
(1) and data acquisition: acquiring the queuing length, the queuing starting position, the number of vehicles and the average speed of a motorcade in a lane by traffic flow equipment, wherein the detected area is the whole lane in a preset detection range; meanwhile, collecting real-time running time data of the traffic signals through traffic signal equipment;
(2) and (3) pretreatment: if the data transmission delay of the traffic flow equipment and the traffic signal equipment is less than the preset range, the data collected by the traffic flow equipment and the traffic signal equipment needs to be removed; in order to synchronize the fusion, the equipment time of the traffic flow equipment and the traffic signal equipment needs to be calibrated into a uniform time;
(3) selecting and defining a characteristic vector: acquiring stage hour flow, stage dissipation rate and stage saturation of the lane;
(4) designing a classifier: and defining an observation matrix x as [ stage hour flow, stage dissipation rate and stage saturation ], and substituting the observation matrix x into an SVM discriminant function to realize state classification.
Further, in the step (1), one end of a detection range is preset, namely the first detection section is close to the intersection vehicle stop line, and the other end of the detection range is preset, namely the second detection section is the lane canalization tail end.
Further, the method for acquiring the flow rate in the stage hour comprises the following steps:
beginning: the stage switching time SplitTime, unit is second;
recording the flow of a first section corresponding to each lane in real time: each of the presence data is a flow; based on the corresponding relation between the traffic flow equipment and the actual road section, matching the lane of the traffic flow equipment with the lane identified by the cost unit, and accumulating the existing data uploaded by the traffic flow equipment to calculate the first section flow of the lane;
recording whether the duration time exceeds the split time and the execution times of each stage are equal: starting from the selected calculation time, when the operation of a certain stage is finished, calculating whether the period of time exceeds the SplitTime, and if so, judging whether the execution times of each stage in the period of time are equal; if the requirements are met, executing the next step, recording the specific duration TotolPhaseLen, and if the requirements are not met, continuing the previous step until the conditions are met;
counting the total flow of each lane in the time corresponding to the section I SplitTime, adding a weight coefficient according to the traffic condition of the lane to adjust the traffic of the lane: counting the total flow of the first section of each lane;
converting lane traffic into flow traffic: based on the corresponding incidence relation of the flow direction lane, converting the lane flow into the flow direction flow DirTatalFlow according to the flow direction lane flow conversion algorithm model fDirTatalFlow;
converting flow direction traffic into stage traffic: based on the stage flow direction corresponding correlation relationship, converting the flow direction flow into the stage flow PhaseTatalFlow according to a stage flow direction flow conversion algorithm model fPhaseTatalFlow;
phase hourly flow was obtained from TotolPhaseLen conversion to hourly flow:
HourFlow＝(PhaseTatalFlow*3600)/(TotolPhaseLen)。
further, the stage dissipation rate obtaining method comprises the following steps:
setting that a cross section of a certain lane corresponding to a certain period of a certain stage passes through ten vehicles, and making a head time distance between the second vehicle and the tenth vehicle and the occupied time of the ten vehicles into a scatter diagram positioned in a two-dimensional rectangular coordinate system, namely an XY coordinate system; the XY coordinate system has two vertical coordinates in the Y direction; the X coordinate is a passing point of each vehicle, the left ordinate is an occupied time ordinate, and the right ordinate is an elapsed time ordinate;
setting a headway function fTHOccupation time function f (x)OG (x), the first order difference between the two is
x is 0,1 … n, when
Summing and averaging to obtain the dissipation ratio
Further, the phase saturation obtaining method comprises the following steps:
recording the number of vehicles in the corresponding lane n1 at the stage operation starting time, recording the number of vehicles in the corresponding lane n2 at the stage operation ending time, recording the stage operation ending time t1 at the time t2 when the number of detected vehicles reaches a certain threshold value, calculating the time difference t to be t2-t1, and calculating and outputting
b＝(w1n1+w2n2–w3t)/(w1N1+w2N2-w3T)*100％，
Wherein N1, N2 and T are different fixed values obtained according to different road canalizations.
Has the advantages that: the invention relates to a space-time analysis technology based on traffic state recognition eigenvector, namely lane-level refined deep fusion analysis is carried out on space traffic flow data and time real-time signal data, and value parameters more fitting with the traffic state are obtained. After the invention is applied to a traffic signal control system, the following advantages are reflected:
(1) the acquired basic source data are comprehensive and rich, and can represent the traffic state better than single basic data;
(2) the characteristic vector for traffic state identification obtained by the space-time analysis technology is more refined and is related to the signal, so that the specific implementation of a signal control strategy can be directly guided.
Drawings
Fig. 1 is a flow chart of the intersection traffic state identification method of the present invention.
Fig. 2 is a schematic view illustrating setting of a lane detection area according to the present invention.
Fig. 3 is a flow chart of the invention for obtaining hourly flow.
Fig. 4 is a scatter diagram of headway occupation time in the present invention.
FIG. 5 is a flow chart of the saturation of the acquisition phase of the present invention.
Detailed Description
The invention discloses a traffic signal coordination control method for automatically adjusting a period and a split green ratio, which comprises the following steps of:
(1) data acquisition, (2) a preprocessing step, (3) a feature vector selection and definition step, and (4) a classifier design step.
Each of the actual operating steps is described below in connection with a particular application in traffic engineering.
(1) Data acquisition
1) The method provides an interface for acquiring spatial data sent by traffic flow equipment, wherein the traffic flow equipment data acquisition mode is a method for detecting an area of a section 1 of an entrance lane space 2, and a schematic diagram of the method is shown in fig. 2. Each lane has two detection sections and a detection area, two detection sections are located about 0.5 meter before the stop line and canalization end respectively, and the area that detects is the whole lane in certain detection range. The two sections are detected to be existence data, the sending triggering condition is that vehicles pass, the sent data has traffic information such as vehicle passing time, vehicle types, speed, occupied time and the like, the detection area is detected to be queuing data, and the sending triggering condition is that the sent data has traffic information such as queuing length, queuing starting position, vehicle number, average speed and the like according to a certain frequency.
The two sections of the stopping line, about 0.5 m ahead and the canalized end are numbered as No. 1 and No. 2 detection sections in sequence.
2) The method also provides an interface to collect traffic signal real-time running time data sent by the traffic signal equipment.
(2) Pretreatment of
1) In order to enable the data obtained by fusion analysis to reflect the road traffic state in time, the data transmission delay of traffic flow equipment and traffic signal equipment is required to be less than a certain range, and if the data transmission delay is greater than the certain range, the data is required to be removed; in order to synchronize the fusion, the time of the traffic flow equipment and the equipment time of the traffic signal equipment need to be corrected for the time of the system to be unified;
2) for time and space data, abnormal data in the time and space data need to be cleaned and removed;
3) the data identified by the mapping cost method, namely the basic data is matched with the space-time data based on the road network data.
(3) Feature vector selection and definition
1) Hourly flow rate
As shown in fig. 3, the phase-hour flow rate acquisition method is as follows.
Beginning: selecting a calculation time, wherein different time granularities can be selected according to different requirements, the method selects a stage switching time, the stage switching time is obtained by obtaining the running state of signal equipment in real time, and the calculation unit is an output unit which calculates according to a constantly moving time window; the split time size is entered in seconds.
Recording the flow of a first section corresponding to each lane in real time: for the raw data of traffic flow equipment, each of the presence data is a flow rate. Based on the corresponding relation between the traffic flow equipment and the actual road section, the lane (lane matching model) identified by the lane matching cost unit of the traffic flow equipment is accumulated and calculated into the first section flow of the lane.
Recording whether the duration time exceeds the split time and the execution times of each stage are equal: starting from the selected calculation time, when the operation of a certain stage is finished, calculating whether the period of time exceeds the SplitTime, and if so, judging whether the execution times of each stage in the period of time are equal. And when the conditions are satisfied, executing the next step, and recording the specific duration TotolPhaseLen, and when the conditions are not satisfied, continuing the previous step until the conditions are satisfied.
And (3) counting the total flow of each lane in the time corresponding to the section I SplitTime, and adding a weight coefficient according to the traffic condition of the lane to adjust the traffic of the lane: and (4) counting the total flow of the first section of each lane, and adjusting the lane weight coefficient by properly adding a lane weight coefficient to the lane under the conditions of turning around or high vehicle occupancy rate.
Converting lane traffic into flow traffic: and converting the lane flow into the flow DirTatalFlow according to the flow conversion algorithm model fDirTatalFlow based on the corresponding incidence relation of the flow lanes.
Converting flow direction traffic into stage traffic: and based on the stage flow direction corresponding correlation relationship, converting the flow direction flow into the stage flow PhaseTatalFlow according to the stage flow direction flow conversion algorithm model fPhaseTatalFlow. Phase hourly flow was obtained from TotolPhaseLen conversion to hourly flow: HourFlow ═ (phasetatalfow 3600)/(TotolPhaseLen).
2) Stage dissipation rate
Assuming that a corresponding lane section passes through ten vehicles during a certain period, the time intervals of the second to tenth vehicles and the occupation time of the ten vehicles are taken as a scatter diagram as shown in fig. 4:
the x coordinate is the passing point of each vehicle, the left ordinate is the ordinate of the occupied time in ms, and the right ordinate is the ordinate of the headway in s.
Setting a headway function fTHOccupation time function f (x)OG (x), the first order difference between the two is
x is 0,1 … n, when
Summing and averaging to obtain the dissipation ratio
3) Stage saturation
As shown in fig. 5, the phase operation starting time records the number N1 of vehicles in the corresponding lane, the phase operation ending time records the number N2 of vehicles in the corresponding lane, the phase operation ending time T1 records the time T2 when the number of detected vehicles reaches a certain threshold, the time difference T is calculated to be T2-T1, and the calculation output b is (w1N1+ w2N2-w 3T)/(w1N1+ w2N2-w3T) × 100%, wherein N1, N2 and T are different fixed values obtained according to channeling at different intersections.
(4) Classifier design
1) Classification
In the national standard GBT 33171-2016 urban traffic operation condition evaluation standard, the operation conditions of traffic are divided into 5 levels of smooth traffic, basic smooth traffic, light congestion, moderate congestion and severe congestion. Referring to the national standard, the method adopts a supervised learning mode to classify the traffic states (not road sections but accurate to the intersection) into five types by using smooth, basically smooth, light congestion, moderate congestion and severe congestion as labels.
2) Learning and training
The traditional statistical pattern recognition method is researched on the premise that the number of samples is enough, and the performance of the proposed method is theoretically guaranteed only when the number of samples tends to infinity. In practice, the number of samples is usually limited. The statistical learning theory which is rapidly developed in recent years is a pattern recognition theory for specially researching small sample learning, a better theoretical framework is established for researching statistical pattern recognition and wider machine learning problems under the condition of limited samples, and a new pattern recognition method, namely a support vector machine, is also developed, so that the small sample learning problem can be better solved.
The support vector machine has many specific advantages in solving the problems of small samples, nonlinearity and high-dimensional pattern recognition, and is decided to be used for learning and classification. The following briefly introduces the principle of pattern learning and classification of the support vector machine in combination with the practical application of the method.
For the method, an observation matrix x is defined as [ stage hour flow, stage dissipation rate, and stage saturation ], a proper kernel function is selected, and the observation matrix x is substituted into an SVM discriminant function to realize state classification.
The SVM method is put forward from an optimal classification surface under the condition of linear classification, and a solution classifier of the SVM method is
Wherein a isiI-1, 2, …, n, is the solution of the following quadratic optimization problem
0≤αi≤C i＝1，2，…，n
Where b can be found by the sample (i.e., the support vector) for which the following equation holds:
for the linear indistinguishable case, SVInstead of explicitly using a specific form of nonlinear transformation, M replaces the vector product operation of the original mode space with a kernel function satisfying the Mercer condition to implement the nonlinear transformation, which essentially transforms the original mode space into a high-dimensional or even infinite-dimensional Hilbert space. If x is transformed nonlinearly, the new characteristic is recorded as
It can be shown that, regardless of the particular form of the transformation, the effect of the transformation on the support vector machine is to put two inner products (x) in the meta-feature spacei*xj) Becomes an inner product in a new space
Note the book
Called kernel function, the support vector machine in the transform space can be written as:
where the coefficient α is the solution to the following optimization problem:
0≤αi≤C i＝1，2，…，n
the coefficient b is obtained from a sample (i.e., a support vector) in which the following equation holds:
for the current commonly used kernel function, through comparison and in combination with the actual concrete conditions of traffic states, as the matrix dimension selected by the method is smaller and the number of samples is general, the method selects a Radial Basis Function (RBF) kernel function as another commonly used linear kernel function which is a special case of the RBF, the problem of linearity does not need to be concerned again when the RBF is used, and in addition, the parameters of the RBF are less than the polynomial kernel function, so that the complexity can be greatly reduced, and the numerical calculation difficulty can be reduced.
The road traffic state recognition system constructed by the method is a multi-class classification problem, and because the traditional support vector machine method only considers the problem of binary classification, an SVM model needs to be expanded to establish a plurality of support vector machine classifiers. At present, two types of methods are mainly used for constructing the SVM multi-class classifier, wherein a direct method is high in calculation complexity and difficult to implement and is only suitable for small-scale problems.
Claims (2)
1. A crossing traffic state identification method based on space-time analysis is characterized by comprising the following steps:
(1) and data acquisition: acquiring the queuing length, the queuing starting position, the number of vehicles and the average speed of a motorcade in a lane by traffic flow equipment, wherein the detected area is the whole lane in a preset detection range; meanwhile, collecting real-time running time data of the traffic signals through traffic signal equipment;
(2) and (3) pretreatment: if the data transmission delay of the traffic flow equipment and the traffic signal equipment is less than the preset range, the data collected by the traffic flow equipment and the traffic signal equipment needs to be removed; in order to synchronize the fusion, the equipment time of the traffic flow equipment and the traffic signal equipment needs to be calibrated into a uniform time;
(3) selecting and defining a characteristic vector: acquiring stage hour flow, stage dissipation rate and stage saturation of the lane;
the method for acquiring the flow in the stage hour comprises the following steps:
beginning: the stage switching time SplitTime, unit is second;
recording the flow of a first section corresponding to each lane in real time: each of the presence data is a flow; based on the corresponding relation between the traffic flow equipment and the actual road section, matching the lane of the traffic flow equipment with the lane identified by the cost unit, and accumulating the existing data uploaded by the traffic flow equipment to calculate the first section flow of the lane;
recording whether the duration time exceeds the split time and the execution times of each stage are equal: starting from the selected calculation time, when the operation of a certain stage is finished, calculating whether the period of time exceeds the SplitTime, and if so, judging whether the execution times of each stage in the period of time are equal; if the requirements are met, executing the next step, recording the specific duration TotolPhaseLen, and if the requirements are not met, continuing the previous step until the conditions are met;
counting the total flow of each lane in the time corresponding to the section I SplitTime, adding a weight coefficient according to the traffic condition of the lane to adjust the traffic of the lane: counting the total flow of the first section of each lane;
converting lane traffic into flow traffic: based on the corresponding incidence relation of the flow direction lane, converting the lane flow into the flow direction flow DirTatalFlow according to the flow direction lane flow conversion algorithm model fDirTatalFlow;
converting flow direction traffic into stage traffic: based on the stage flow direction corresponding correlation relationship, converting the flow direction flow into the stage flow PhaseTatalFlow according to a stage flow direction flow conversion algorithm model fPhaseTatalFlow;
phase hourly flow was obtained from TotolPhaseLen conversion to hourly flow:
HourFlow＝(PhaseTatalFlow*3600)/(TotolPhaseLen)；
the method for acquiring the stage dissipation rate comprises the following steps:
setting that a cross section of a certain lane corresponding to a certain period of a certain stage passes through ten vehicles, and making a head time distance between the second vehicle and the tenth vehicle and the occupied time of the ten vehicles into a scatter diagram positioned in a two-dimensional rectangular coordinate system, namely an XY coordinate system; the XY coordinate system has two vertical coordinates in the Y direction; the X coordinate is a passing point of each vehicle, the left ordinate is an occupied time ordinate, and the right ordinate is an elapsed time ordinate;
setting a headway function fTHOccupation time function f (x)OG (x), the first order difference between the two is
When in use
Summing and averaging to obtain the dissipation ratio
The method for acquiring the stage saturation comprises the following steps:
recording the number of vehicles in the corresponding lane n1 at the stage operation starting time, recording the number of vehicles in the corresponding lane n2 at the stage operation ending time, recording the stage operation ending time t1 at the time t2 when the number of detected vehicles reaches a certain threshold value, calculating the time difference t to be t2-t1, and calculating and outputting
b＝(w1n1+w2n2-w3t)/(w1N1+w2N2-w3T)*100％，
Wherein N1, N2 and T are different fixed values obtained according to different road canalizations;
(4) designing a classifier: and defining an observation matrix x as [ stage hour flow, stage dissipation rate and stage saturation ], and substituting the observation matrix x into an SVM discriminant function to realize state classification.
2. The intersection traffic state recognition method according to claim 1, wherein in the step (1), one end of the preset detection range, namely the first detection section, is close to a vehicle stop line at the intersection, and the other end of the preset detection range, namely the second detection section, is a lane canalization end.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811554437.4A CN109754599B (en) | 2018-12-19 | 2018-12-19 | Crossing traffic state identification method based on space-time analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811554437.4A CN109754599B (en) | 2018-12-19 | 2018-12-19 | Crossing traffic state identification method based on space-time analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109754599A CN109754599A (en) | 2019-05-14 |
| CN109754599B true CN109754599B (en) | 2022-01-04 |
Family
ID=66403723
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811554437.4A Active CN109754599B (en) | 2018-12-19 | 2018-12-19 | Crossing traffic state identification method based on space-time analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109754599B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20160062880A (en) * | 2014-11-26 | 2016-06-03 | 휴앤에스(주) | road traffic information management system for g using camera and radar |
| CN108171365A (en) * | 2017-12-15 | 2018-06-15 | 南京理工大学 | A kind of traffic status prediction method based on improvement SVM algorithm |
| CN108417039A (en) * | 2018-05-07 | 2018-08-17 | 大连理工大学 | A kind of signalized intersections transport need evaluation method that traffic flow composition influences |
| CN108492568A (en) * | 2018-04-25 | 2018-09-04 | 南京邮电大学 | A kind of Short-time Traffic Flow Forecasting Methods based on space-time characterisation analysis |
-
2018
- 2018-12-19 CN CN201811554437.4A patent/CN109754599B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20160062880A (en) * | 2014-11-26 | 2016-06-03 | 휴앤에스(주) | road traffic information management system for g using camera and radar |
| CN108171365A (en) * | 2017-12-15 | 2018-06-15 | 南京理工大学 | A kind of traffic status prediction method based on improvement SVM algorithm |
| CN108492568A (en) * | 2018-04-25 | 2018-09-04 | 南京邮电大学 | A kind of Short-time Traffic Flow Forecasting Methods based on space-time characterisation analysis |
| CN108417039A (en) * | 2018-05-07 | 2018-08-17 | 大连理工大学 | A kind of signalized intersections transport need evaluation method that traffic flow composition influences |
Non-Patent Citations (2)
| Title |
|---|
| 于荣 等.基于支持向量机的城市道路交通状态模式识别研究.《交通运输系统工程与信息》.2013,第13卷(第1期),第130-136页. * |
| 基于支持向量机的城市道路交通状态模式识别研究;于荣 等;《交通运输系统工程与信息》;20130228;第13卷(第1期);第130-136页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109754599A (en) | 2019-05-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106355907B (en) | Signalized crossing queue length real-time estimation method based on track of vehicle | |
| CN102819952B (en) | Method for detecting illegal lane change of vehicle based on video detection technique | |
| WO2022206974A1 (en) | Roadside sensing unit-based static and non-static object point cloud recognition method | |
| WO2022252398A1 (en) | Ship trajectory feature point extraction-based spatio-temporal dp method | |
| CN109902806A (en) | Method is determined based on the noise image object boundary frame of convolutional neural networks | |
| Fang et al. | FTPG: A fine-grained traffic prediction method with graph attention network using big trace data | |
| CN105608505A (en) | Cellular signaling data based track traffic travel mode identification method for resident | |
| CN103903441B (en) | Road traffic state distinguishing method based on semi-supervised learning | |
| CN105513349A (en) | Double-perspective learning-based mountainous area highway vehicle event detection method | |
| CN106920403A (en) | A kind of single-point self-adaptation control method based on array radar | |
| CN105354542A (en) | Method for detecting abnormal video event in crowded scene | |
| CN104951764A (en) | Identification method for behaviors of high-speed vehicle based on secondary spectrum clustering and HMM (Hidden Markov Model)-RF (Random Forest) hybrid model | |
| CN103839415A (en) | Traffic flow and occupation ratio information acquisition method based on road surface image feature identification | |
| CN103164697B (en) | The licence plate recognition method of a kind for the treatment of time and accuracy of identification self-adaptation | |
| CN111882858A (en) | Method for predicting queuing length of expressway abnormal events based on multi-source data | |
| CN114023062A (en) | Traffic flow information monitoring method based on deep learning and edge calculation | |
| CN114372503A (en) | Cluster vehicle motion trail prediction method | |
| CN106846808B (en) | A kind of vehicle parking based on license plate data time number calculating method | |
| CN109859475B (en) | Intersection signal control method, device and system based on DBSCAN density clustering | |
| CN103577804A (en) | Abnormal human behavior identification method based on SIFT flow and hidden conditional random fields | |
| CN109584564B (en) | Method, system and device for implementing and optimizing multi-device communication control scheme | |
| CN109754599B (en) | Crossing traffic state identification method based on space-time analysis | |
| CN116704750B (en) | Traffic state identification method based on clustering algorithm, electronic equipment and medium | |
| CN103065152B (en) | The recognition methods of digital dock in a kind of video | |
| CN105069407B (en) | A kind of magnitude of traffic flow acquisition methods based on video |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |