CN113096397A - Traffic jam analysis method based on millimeter wave radar and video detection - Google Patents
Traffic jam analysis method based on millimeter wave radar and video detection Download PDFInfo
- Publication number
- CN113096397A CN113096397A CN202110348509.5A CN202110348509A CN113096397A CN 113096397 A CN113096397 A CN 113096397A CN 202110348509 A CN202110348509 A CN 202110348509A CN 113096397 A CN113096397 A CN 113096397A
- Authority
- CN
- China
- Prior art keywords
- target
- vehicle
- road
- millimeter wave
- wave radar
- 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
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/017—Detecting movement of traffic to be counted or controlled identifying vehicles
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/04—Detecting movement of traffic to be counted or controlled using optical or ultrasonic detectors
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/052—Detecting movement of traffic to be counted or controlled with provision for determining speed or overspeed
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a traffic jam analysis method based on millimeter wave radar and video detection. According to the invention, a video monitoring camera is used for acquiring road images in a field of view, identifying the motion condition of vehicles, measuring data such as the number of vehicles and lanes in the field of view, detecting the positions of the vehicles and the running speed of the vehicles by using a millimeter wave radar, calculating a road congestion index according to the acquired traffic flow data, obtaining a quantitative parameter of congestion degree, and further analyzing the traffic congestion condition. The invention analyzes the traffic jam degree according to the traffic flow data, and intelligently and quickly evaluates the road condition in real time, thereby being beneficial to guiding evacuation traffic jam, improving road conditions and improving the traveling efficiency of citizens.
Description
Technical Field
The invention belongs to the field of traffic road condition supervision, and particularly relates to a traffic jam analysis technology based on radar signals and machine vision.
Background
At present, many road traffic jam analysis technologies based on video images exist at home and abroad, and the technologies can be divided into two parts, namely traffic flow parameter extraction based on the video images and traffic jam analysis based on the traffic flow parameters. A great deal of research is respectively carried out on two parts of contents at home and abroad, but the two parts of contents are not combined for a long time, and more traffic flow parameters for evaluating the traffic jam degree are extracted from videos shot by a fixed traffic camera or extracted from sensor detection equipment such as a coil or simulation of computer software.
Based on traffic video congestion analysis, foreign research is mainly focused on traffic event monitoring. The method mainly comprises the following steps: analyzing a traffic event by extracting a background in the video and a motion track of a vehicle; tracking the vehicle in the video by adopting Kalman filtering so as to realize the detection of the event; by distinguishing the image data into a high-order model and a low-order processing model, the vehicle detection method based on rule reasoning is provided for the monitoring system. The Zhejiang university in China firstly provides a video-based traffic incident detection technology, data mining is carried out on traffic videos by adopting an ART2 neural network algorithm, traffic incidents are automatically detected, and certain achievements are obtained. The Harbin university of industry has also developed a set of information collection systems based on traffic video. The achievement Vload detection system detects road events by analyzing video data and has an early warning function. In short, although such systems can better complete part of video detection functions, most of the detection systems are rarely popularized due to high maintenance cost or insignificant actual use effect.
Most of the existing traffic jam analysis methods basically begin from the collection of traffic information, and obtain traffic rules by analyzing and summarizing the acquired traffic state parameters such as traffic flow, average travel speed and occupancy, so as to establish a relation model between the traffic state parameters and road jam, further obtain a traffic jam analysis model, and finally obtain the degree of road jam. With regard to a relation model between traffic flow parameter indexes and traffic congestion, many researchers at home and abroad contribute various theoretical models. Due to the limitation of traffic data acquisition technology, a relation model between traffic flow parameter indexes and congestion is subject to the development from a speed-density model to a speed-flow model, and further to the flow-density model. However, these three models have various problems and their use is limited by road maximum load and road class. In recent years, researchers at home and abroad propose various traffic jam analysis models according to actual traffic operation characteristics. Foreign scholars extract 5 parameters such as vehicle speed and vehicle flow from the image and then establish a relation model between basic traffic parameters and road congestion by using a hidden Markov model, thereby realizing the evaluation of road service level. While domestic scholars analyze a time sequence chart of road occupancy and occupancy variance, consider that congestion is easily caused when the values of the two parameters are high, and introduce an absolute value of the occupancy in order to observe whether the traffic state changes, and further experiments find that the road congestion is considered when the absolute value of the occupancy changes slowly or does not change basically in a long time, so as to establish the relationship between the three traffic flow parameters and the congestion. The models are complex and difficult to be intuitively understood by non-professionals, so that an intuitive, reasonable and efficient congestion analysis model is needed.
In summary, the traffic flow data collection mode mainly involved is limited to the traffic video collection, and with the development of the detection technology, the characteristics of all weather, wide action area and the like of the road radar detection equipment make the detection equipment well make up for the defect of single video detection, so that the advantages of low cost, high cost performance, high efficiency and the like brought by the detection technology are more and more emphasized at home and abroad. The invention aims to integrate traffic flow parameter extraction and traffic congestion analysis based on the traffic flow parameters, extract vehicle parameters by using a millimeter wave radar and video detection mode, and perform congestion analysis by adopting a scientific and feasible traffic congestion analysis method so as to better guide and improve traffic road conditions.
Disclosure of Invention
The method integrates traffic flow parameter extraction and traffic jam analysis based on the traffic flow parameters, and compared with the method for detecting vehicles by a single video, the method for detecting the traffic jam of the vehicle by the millimeter wave radar and the video acquires the vehicle operation parameters in a linkage mode, identifies the vehicle motion state (including static state or running at a certain speed), and analyzes the traffic road condition by a scientific and efficient jam analysis method based on the acquired traffic flow parameters so as to provide guidance for improving the traffic jam.
The technical scheme provided by the invention is as follows: the method comprises the steps of acquiring road images in a view field by using a video monitoring camera, identifying the motion condition of vehicles, measuring data such as the number of vehicles, lanes and the like in the view field, detecting the positions of the vehicles and the running speeds of the vehicles by using a millimeter wave radar, calculating road congestion indexes according to the acquired traffic flow data, obtaining quantitative parameters of congestion degrees, and further analyzing the traffic congestion conditions. The method specifically comprises the following steps:
step 1, acquiring road information and vehicle quantity information by using a video image, wherein the information comprises the number of lanes in a view field, the length of each lane, the number of vehicles in the view field and the length of each vehicle, calculating the space occupancy of the lanes according to the length of each vehicle and the length of the lanes, and calculating the vehicle group density according to the number of lanes, the length of each lane and the number of vehicles;
step 2, acquiring position information and speed information of the vehicle by using a road vehicle detection system based on a millimeter wave radar, wherein the position information of the vehicle is used for positioning, and the speed information is used for calculating the speed of a vehicle group;
step 3, a road congestion index is constructed for measuring the traffic congestion level, and the calculation formula of the road congestion index is as follows:
the system comprises a vehicle cluster, a remote control unit (RCI), a vehicle cluster speed acquisition unit, a vehicle cluster density acquisition unit, a vehicle cluster speed acquisition unit and a vehicle cluster speed acquisition unit, wherein VS represents the vehicle cluster speed, VD represents the vehicle cluster density, p represents a set parameter, RCI is continuous data, and smaller values represent that a road;
setting a congestion state threshold value R1, setting a threshold value T1 for the duration of the vehicle state of each lane in the visual field, and if the RCI value reaches the congestion state threshold value R1 and the duration exceeds the threshold value T1, determining that the road section is congested;
and 4, outputting the lane space occupancy of the road section in the monitoring field, the corresponding traffic jam level and the position information of the jammed vehicle.
Further, the relationship between the road congestion index and the traffic congestion level is as follows,
further, the lane space occupancy O is a ratio of a total length of the vehicle running on the lane to the length L of the lane at a certain time t, and assuming that n vehicles run together without considering an influence of the volume of the vehicle, a length of an i-th vehicle is LiAnd then the space occupancy of the lane is as follows:
the space occupancy of the lane is an important index for measuring whether the road is fully utilized and the degree of congestion, and has important significance for traffic management.
Further, the vehicle density group VD is a ratio of the number of vehicles traveling on the road to the road area, and assuming that the influence of the vehicle volume is not considered, n lanes are provided in total, the road length is L, and m vehicles are traveling in total, the vehicle group density is:
further, the vehicle group velocity VS is an average velocity of vehicles traveling on a road, assuming that the road is sharedM vehicles are detected, each vehicle is tracked after the vehicles are detected, and the running speed V of each vehicle is detectediThen, the calculation formula of the group speed of the vehicles is:
further, the road vehicle detection system of the millimeter wave radar comprises a millimeter wave radar system and a radar data processing system, wherein the millimeter wave radar system is arranged on a cross bar of a traffic signal lamp, actively transmits electromagnetic wave signals, provides three data information of a relative distance, a relative speed and an azimuth angle of a detected target by calculating Doppler frequency shift of received signals and transmitted signals, outputs the three data information to the data processing system, completes signal filtering by the data processing system, and finally outputs position information and speed information of a vehicle; the filtering comprises two steps: the method comprises effective target primary selection and target validity check, wherein the effective target primary selection is to eliminate false targets which are not in a road area by utilizing a polar coordinate system operation rule, and the target validity check is to utilize a Kalman filtering algorithm to calculate the continuity of the appearance and movement of targets in adjacent periods so as to eliminate the interference of noise data.
Further, the specific implementation manner of the effective target initial selection is as follows;
dividing the effective area of the radar detection target, wherein the effective area is formed by dividing x1 ═ A1,x2＝A2,y1＝B1,y2＝B2A rectangle enclosed therein, wherein A1，A2，B1，B2Boundary values of four sides of the representation rectangle under a rectangular coordinate system with the millimeter wave radar as an origin are obtained by prior knowledge;
the formula for judging the validity of the target, namely whether the target is in the road area is as follows:
wherein r is the relative distance between the target and the radar, α is the azimuth angle between the target and the radar, and if r and α of a certain target conform to the formula (4), it can be considered to be valid.
Further, if the position and speed information of the target to be tracked and the position and speed information obtained by using the Kalman filtering in the continuous N periods conform to a difference limit, the target is considered to be an effective target, and based on the Kalman filtering, the optimal information of the target in the Nth period is obtained by using the Kalman filtering result in the (N-1) th period and the observation result in the Nth period.
Further, the specific implementation steps of the target validity check are as follows;
setting a period threshold value N;
determining a target to be tracked, and selecting a target newly appearing in the period as the target to be tracked in the 1 st period;
calculating a predicted state vector of the target, wherein the state vector describing the target to be tracked in the nth period needs to be obtained, the state vector comprises position information and speed information, and N is more than or equal to 1 and less than or equal to N-1: xn|n＝[xn|n,yn|n,vxn|n,vyn|n]TAnd its covariance Pn|nWherein X represents a state vector, X represents the abscissa of the target to be tracked under a rectangular coordinate system with the millimeter wave radar as the origin, y represents the ordinate, and v represents the position of the target to be trackedxRepresenting the component, v, of the velocity vector of the object to be tracked on the abscissa axisyThe subscript n | n represents a value obtained by calculation through a Kalman filtering method after an observation value of the nth period is added, and if the subscript is n | n-1, the value is a value obtained by prediction through the Kalman filtering method by utilizing a state vector of the (n-1) th period;
if the period ordinal number n is 1, converting the position coordinate of the target to be tracked from a polar coordinate system to a rectangular coordinate system to obtain a position observed value, converting the position coordinate and the velocity vector of the target to be tracked from the polar coordinate system to the rectangular coordinate system by xy orthogonal decomposition to obtain a velocity observed value, wherein the formula is as follows:
wherein r is the relative distance between the target and the radar, α is the azimuth angle between the target and the radar, v is the relative velocity between the target and the radar, subscript 1 represents the cycle number, then X1|1＝[x1|1，y1|1，vx1|1，vy1|1]T＝[x1,y1,vx1,vy1]T，P1|1Supplied by millimeter wave radar manufacturers;
if the cycle number n ≠ 1, it means that the result X calculated using Kalman filtering has been obtained in the previous cyclen|n，Pn|n；
Assuming that the target approximately moves linearly at a constant speed in one period, the predicted state vector of the target in the (n +1) th period can be predicted by equation (7):
namely, it is
Wherein T is a radar scanning period, xn+1|n、yn+1|n、vxn+1|n、vyn+1|nThe predicted state vector of the target for the (n +1) th cycle calculated from the nth cycle
The covariance is calculated by equation (8):
Pn+1|n＝FPn|nFT+Qn (8)
wherein QnA Gaussian covariance matrix with a mean value of zero;
judging whether the two adjacent periods observe the same target, comparing the predicted state vector of the target with the state vector obtained by actual observation, and setting the actual state vector of the target observed in the (n +1) th period as: xn+1＝[xn+1,yn+1,vxn+1,vyn+1]TThen the comparison formula is as follows:
wherein, Δ x, Δ y, Δ vx，ΔvyIs an error margin set by man, if Xn+1And
if equation (9) is satisfied, the target observed in the (n +1) th cycle and the target observed in the nth cycle are considered to be the same target; if the target is the same target, performing the next step, otherwise, considering that the target is lost, namely the target is noise;
calculating a Kalman filtering result, wherein the Kalman filtering result of the (n +1) th period is as follows:
Pn+1|n+1＝Pn+1|n-KnPn+1|n (11)
Kn＝Pn+1|n(Pn+1|n+Rn+1),Rn+1＝P1|1 (12)
wherein, KnFor Kalman gain, Rn+1The covariance matrix observed in the (n +1) th period is R because the error of each observation of the same millimeter wave radar is approximately considered to be equaln+1＝P1|1；
Sixthly, repeating the steps until the target disappears or disappearsIf the period number reaches N, the target is considered to be noise and the tracking is abandoned if the target disappears before the period number reaches N; if the cycle number reaches N, i.e. N +1 equals N, then X is outputn+1|n+1。
Compared with the prior art, the invention has the following advantages and beneficial effects:
(1) the integration of the millimeter wave radar and the video camera can better solve the problems of redundancy and errors of a single sensor when information is used for acquiring traffic flow data, and the robustness and the accuracy of the system are improved.
(2) The video camera detects traffic data and is easily influenced by weather, the action range is limited, and the millimeter wave radar detection equipment well makes up the defect of single video detection due to the characteristics of all weather, wide action range and the like, and has the advantages of low cost, high cost performance, high efficiency and the like.
(3) The road congestion index is used for quantifying the traffic congestion degree, and the acquired information such as vehicle density and speed is fully utilized, so that the method has the characteristics of science, conciseness and intuition.
(4) The traffic jam degree is analyzed according to the traffic flow data, and the road condition is evaluated intelligently and quickly in real time, so that the traffic jam can be guided to evacuate, the road condition is improved, and the citizen travel efficiency is improved.
Drawings
Fig. 1 is a flow chart of traffic flow data acquisition and traffic congestion analysis implemented by the present invention.
Fig. 2 is a schematic view of a first step of traffic flow data acquisition according to the present invention.
Detailed Description
The technical solution of the present invention is further explained with reference to the drawings and the embodiments.
As shown in fig. 1, the present invention provides a traffic congestion analysis method based on millimeter wave radar and video detection, including the following steps:
step 1, acquiring road information and vehicle quantity information by using a video image, wherein the information comprises the number of lanes in a view field, the length of each lane, the number of vehicles in the view field and the length of each vehicle, calculating the space occupancy of the lanes according to the length of each vehicle and the length of the lanes, and calculating the vehicle group density according to the number of lanes, the length of each lane and the number of vehicles;
acquiring vehicle image information in a view field l through cameras arranged around a monitored road, extracting vehicle characteristic profiles, calibrating the positions of vehicles in the view field, and continuously tracking; and converting the lane real-time traffic data information into lane real-time traffic data information, and transmitting the lane real-time traffic data information to a traffic information processing server, wherein the lane real-time traffic data information comprises the number of lanes in a view field, the length of each lane, the number of vehicles in the view field and the length of each vehicle.
The lane space occupancy rate O refers to the ratio of the total length of the vehicles running on the lane to the length L of the lane at a certain time t, n vehicles run together under the assumption that the influence of the volume of the vehicles is not considered, and the length of the ith vehicle is LiAnd then the space occupancy of the lane is as follows:
the space occupancy of the lane is an important index for measuring whether the road is fully utilized and the degree of congestion, and has important significance for traffic management.
The vehicle density group VD is a ratio of the number of vehicles running on the road to the road area, and assuming that the influence of the vehicle volume is not considered, n lanes are provided in total, the road length is L, and m vehicles are running in total, then the vehicle group density is:
step 2, acquiring position information and speed information of the vehicle by using a road vehicle detection system based on a millimeter wave radar, wherein the position information of the vehicle is used for positioning, and the speed information is used for calculating the speed of a vehicle group;
the road vehicle detection system of the millimeter wave radar comprises a millimeter wave radar system and a radar data processing system, wherein the millimeter wave radar system is arranged on a cross bar of a traffic signal lamp, actively transmits electromagnetic wave signals, provides three data information of relative distance, relative speed and azimuth angle of a detection target by calculating Doppler frequency shift of received signals and transmitted signals, then outputs the three data information to the data processing system, and the data processing system completes signal filtering and finally outputs position information and speed information of a vehicle; the filtering comprises two steps: the method comprises effective target primary selection and target validity check, wherein the effective target primary selection is to eliminate false targets which are not in a road area by utilizing a polar coordinate system operation rule, and the target validity check is to utilize a Kalman filtering algorithm to calculate the continuity of the appearance and movement of targets in adjacent periods so as to eliminate the interference of noise data.
In addition to the valid target on the road, the target acquired by the millimeter wave radar may have two kinds of interference targets: targets outside a road area may be pedestrians or other obstacles in motion, and are considered as interference targets due to the fact that the targets exceed the action field of the technology; ② false objects, i.e. noise signals, which appear and disappear in a very short time. If the two interference targets cannot be accurately eliminated, the accuracy of the detection result is affected. And now, target screening is carried out by combining two modes of effective target primary selection and target effectiveness test so as to improve the quality of the obtained target information.
(1) Efficient object primary selection
The detection range of the millimeter wave radar is wider, if the original detection range of the millimeter wave radar is reserved, the detection and the processing of the non-road motor vehicle target can be increased while more target information is detected, the processing time is increased, and the accuracy of an output result is influenced.
As shown in FIG. 2, the effective area of the radar detection target is divided, and the effective area is defined by x1＝A1,x1＝A2,y1＝B1,y1＝B2A rectangle enclosed therein, wherein A1，A2，B1，B2The boundary value of four sides of the representation rectangle under the rectangular coordinate system with the millimeter wave radar as the origin is determined a prioriKnowledge derived (either manually measured or output identified by a road line detection algorithm).
The formula for judging the validity of the target, namely whether the target is in the road area is as follows:
wherein r is the relative distance between the target and the radar, and α is the azimuth angle between the target and the radar. If r and α of a certain target match equation (4), it can be considered valid (in the road region).
(2) Target validity check
The validity verification of the target data information mainly aims at false targets with a flash characteristic generated due to the instability of the operation of the millimeter wave radar.
The principle of target validity check is as follows: and if the errors between the position and speed information of the target to be tracked and the position and speed information obtained by Kalman filtering in N continuous periods conform to a difference limit, the target is considered to be an effective target (non-noise), and based on the Kalman filtering, the optimal information of the target in the Nth period is obtained by utilizing the Kalman filtering result in the (N-1) th period and the observation result in the Nth period.
Setting a period threshold value N.
Determining the target to be tracked. In the 1 st period, a target newly appearing in the period is selected as a target to be tracked.
And calculating the predicted state vector of the target. Firstly, a state vector (the state comprises a position and a speed) for describing a target to be tracked in the nth period (N is more than or equal to 1 and less than or equal to N-1) needs to be obtained: xn|n＝[xn|n,yn|n,vxn|n,vyn|n]TAnd its covariance Pn|nWherein X represents a state vector, X represents the abscissa of the target to be tracked under a rectangular coordinate system with the millimeter wave radar as the origin, y represents the ordinate, and v represents the position of the target to be trackedxRepresenting the component, v, of the velocity vector of the object to be tracked on the abscissa axisyRepresenting components on the ordinate axis, the subscript n representing additionA value calculated by the kalman filter method after the observation value of the nth cycle is obtained (if the subscript is n | n-1, this indicates that the value is a value predicted by the kalman filter method using the state vector of the (n-1) th cycle).
If the period ordinal number n is 1, converting the position coordinate of the target to be tracked from a polar coordinate system to a rectangular coordinate system to obtain a position observed value, converting the position coordinate and the velocity vector of the target to be tracked from the polar coordinate system to the rectangular coordinate system by xy orthogonal decomposition to obtain a velocity observed value, wherein the formula is as follows:
wherein r is the relative distance between the target and the radar, α is the azimuth angle between the target and the radar, v is the relative velocity between the target and the radar, subscript 1 represents the cycle number, then X1|1＝[x1|1,y1|1,vx1|1,vy1|1]T＝[x1,y1,vx1,vy1]T，P1|1Supplied by millimeter wave radar manufacturers.
If the cycle number n ≠ 1, it means that the result X calculated using Kalman filtering has been obtained in the previous cyclen|n，Pn|n。
Assuming that the target can approximately perform a uniform linear motion in the interval of one cycle, the predicted state vector of the target in the (n +1) th cycle can be predicted by equation (7):
namely, it is
Where T is the radar scan period, typically 50ms, xn+1|n、yn+1|n、vxn+1|n、vyn+1|nThe predicted state vector of the target for the (n +1) th cycle calculated from the nth cycle
The covariance can be calculated by equation (8):
Pn+1|n＝FPn|nFT+Qn (8)
wherein QnIs a gaussian covariance matrix with a mean of zero.
Judging whether the two adjacent periods are observed to be the same target or not. Comparing the predicted state vector of the target with the state vector obtained by actual observation, and setting the actual state vector of the target observed in the (n +1) th period as follows: xn+1＝[xn+1,yn+1,vxn+1,vyn+1]TThen the comparison formula is as follows:
wherein, Δ x, Δ y, Δ vx，ΔvyIs an error margin set by human. If Xn+1And
the object observed in the (n +1) th cycle and the same object observed in the nth cycle are considered to be the same object if equation (9) is satisfied. If the target is the same, the next step is carried out, otherwise, the target is considered to be lost, namely the target is noise.
And fifthly, calculating a Kalman filtering result. The kalman filtering result of the (n +1) th cycle is:
Pn+1|n+1＝Pn+1|n-KnPn+1|n (11)
Kn＝Pn+1|n(Pn+1|n+Rn+1),Rn+1＝P1|1 (12)
wherein, KnFor Kalman gain, Rn+1The covariance matrix observed in the (n +1) th period has R because the error of each observation of the same millimeter wave radar can be approximately considered to be equaln+1＝P1|1.
Sixthly, repeating the step three to the fifth until the target disappears or the period number reaches N. If the target disappears before the cycle number reaches N, the target is considered to be noise, and the tracking is abandoned. If the cycle number reaches N, i.e. N +1 equals N, then X is outputn+1|n+1。
The vehicle group speed VS is the average speed of the vehicles running on the road, assuming that there are m vehicles on the road, tracking each vehicle after detecting the vehicles, and detecting the running speed V of each vehicleiThen, the calculation formula of the group speed of the vehicles is:
step 3, a road congestion index is constructed for measuring the traffic congestion level, and the calculation formula of the road congestion index is as follows:
the system comprises a vehicle cluster, a remote control unit (RCI), a vehicle cluster speed acquisition unit, a vehicle cluster density acquisition unit, a vehicle cluster speed acquisition unit and a vehicle cluster speed acquisition unit, wherein VS represents the vehicle cluster speed, VD represents the vehicle cluster density, p represents a set parameter, RCI is continuous data, and smaller values represent that a road;
setting a congestion state threshold value R1, setting a threshold value T1 for the duration of the vehicle state of each lane in the visual field, and if the RCI value reaches the congestion state threshold value R1 and the duration exceeds the threshold value T1, determining that the road section is congested;
the road congestion index is related to the traffic congestion level as follows,
and 4, outputting the lane space occupancy of the road section in the monitoring field, the corresponding traffic jam level and the position information of the jammed vehicle. Through the radar and video linkage detection system and congestion modeling analysis, the specific place where the congested traffic flow occurs can be judged, and the analysis result is output to complete the traffic congestion analysis function.
The specific embodiments described herein are merely illustrative of the spirit of the invention. Various modifications or additions may be made to the described embodiments or alternatives may be employed by those skilled in the art without departing from the spirit or ambit of the invention as defined in the appended claims.
Claims (9)
1. The traffic jam analysis method based on the millimeter wave radar and the video detection is characterized by comprising the following steps of:
step 1, acquiring road information and vehicle quantity information by using a video image, wherein the information comprises the number of lanes in a view field, the length of each lane, the number of vehicles in the view field and the length of each vehicle, calculating the space occupancy of the lanes according to the length of each vehicle and the length of the lanes, and calculating the vehicle group density according to the number of lanes, the length of each lane and the number of vehicles;
step 2, acquiring position information and speed information of the vehicle by using a road vehicle detection system based on a millimeter wave radar, wherein the position information of the vehicle is used for positioning, and the speed information is used for calculating the speed of a vehicle group;
step 3, a road congestion index is constructed for measuring the traffic congestion level, and the calculation formula of the road congestion index is as follows:
the system comprises a vehicle cluster, a remote control unit (RCI), a vehicle cluster speed acquisition unit, a vehicle cluster density acquisition unit, a vehicle cluster speed acquisition unit and a vehicle cluster speed acquisition unit, wherein VS represents the vehicle cluster speed, VD represents the vehicle cluster density, p represents a set parameter, RCI is continuous data, and smaller values represent that a road;
setting a congestion state threshold value R1, setting a threshold value T1 for the duration of the vehicle state of each lane in the visual field, and if the RCI value reaches the congestion state threshold value R1 and the duration exceeds the threshold value T1, determining that the road section is congested;
and 4, outputting the lane space occupancy of the road section in the monitoring field, the corresponding traffic jam level and the position information of the jammed vehicle.
2. The traffic congestion analysis method based on millimeter wave radar and video detection as claimed in claim 1, wherein: the road congestion index is related to the traffic congestion level as follows,
3. the traffic congestion analysis method based on millimeter wave radar and video detection as claimed in claim 1, wherein: the lane space occupancy rate O is the ratio of the total length of the vehicles running on the lane to the length L of the lane at a certain time t, n vehicles run together under the assumption that the influence of the volume of the vehicles is not considered, and the length of the ith vehicle is LiAnd then the space occupancy of the lane is as follows:
the space occupancy of the lane is an important index for measuring whether the road is fully utilized and the degree of congestion, and has important significance for traffic management.
4. The traffic congestion analysis method based on millimeter wave radar and video detection as claimed in claim 1, wherein: the vehicle density group VD is a ratio of the number of vehicles running on the road to the road area, and assuming that the influence of the vehicle volume is not considered, there are n lanes in total, the road length is L, and m vehicles are running in total, then the vehicle group density is:
5. the traffic congestion analysis method based on millimeter wave radar and video detection as claimed in claim 1, wherein: the vehicle group speed VS is the average speed of the vehicles running on the road, assuming that there are m vehicles on the road, tracking each vehicle after detecting the vehicles, and detecting the running speed V of each vehicleiThen, the calculation formula of the group speed of the vehicles is:
6. the traffic congestion analysis method based on millimeter wave radar and video detection as claimed in claim 1, wherein: the road vehicle detection system of the millimeter wave radar comprises a millimeter wave radar system and a radar data processing system, wherein the millimeter wave radar system is arranged on a cross bar of a traffic signal lamp, the millimeter wave radar system actively transmits electromagnetic wave signals, three data information of relative distance, relative speed and azimuth angle of a detection target are provided by calculating Doppler frequency shift of received signals and transmitted signals, then the three data information are output to the data processing system, the data processing system completes signal filtering, and finally position information and speed information of a vehicle are output; the filtering comprises two steps: the method comprises effective target primary selection and target validity check, wherein the effective target primary selection is to eliminate false targets which are not in a road area by utilizing a polar coordinate system operation rule, and the target validity check is to utilize a Kalman filtering algorithm to calculate the continuity of the appearance and movement of targets in adjacent periods so as to eliminate the interference of noise data.
7. The traffic congestion analysis method based on millimeter wave radar and video detection as claimed in claim 6, wherein: the specific implementation manner of the effective target initial selection is as follows;
dividing the effective area of the radar detection target, wherein the effective area is formed by dividing x1 ═ A1,x2＝A2,y1＝B1,y2＝B2A rectangle enclosed therein, wherein A1，A2，B1，B2Boundary values of four sides of the representation rectangle under a rectangular coordinate system with the millimeter wave radar as an origin are obtained by prior knowledge;
the formula for judging the validity of the target, namely whether the target is in the road area is as follows:
wherein r is the relative distance between the target and the radar, α is the azimuth angle between the target and the radar, and if r and α of a certain target conform to the formula (4), it can be considered to be valid.
8. The traffic congestion analysis method based on millimeter wave radar and video detection as claimed in claim 6, wherein: and if the errors between the position and speed information of the target to be tracked and the position and speed information obtained by using Kalman filtering in N continuous periods accord with a difference limit, the target is considered to be an effective target, and based on the Kalman filtering, the optimal information of the target in the N period is obtained by using the Kalman filtering result in the (N-1) th period and the observation result in the N period.
9. The traffic congestion analysis method based on millimeter wave radar and video detection as claimed in claim 6, wherein: the specific implementation steps of the target validity check are as follows;
setting a period threshold value N;
determining a target to be tracked, and selecting a target newly appearing in the period as the target to be tracked in the 1 st period;
calculating a predicted state vector of the target, wherein the state vector describing the target to be tracked in the nth period needs to be obtained, the state vector comprises position information and speed information, and N is more than or equal to 1 and less than or equal to N-1: xn|n＝[xn|n,yn|n,vxn|n,vyn|n]TAnd its covariance Pn|nWherein X represents a state vector, X represents the abscissa of the target to be tracked under a rectangular coordinate system with the millimeter wave radar as the origin, y represents the ordinate, and v represents the position of the target to be trackedxRepresenting the component, v, of the velocity vector of the object to be tracked on the abscissa axisyThe subscript n | n represents a value obtained by calculation through a Kalman filtering method after an observation value of the nth period is added, and if the subscript is n | n-1, the value is a value obtained by prediction through the Kalman filtering method by utilizing a state vector of the (n-1) th period;
if the period ordinal number n is 1, converting the position coordinate of the target to be tracked from a polar coordinate system to a rectangular coordinate system to obtain a position observed value, converting the position coordinate and the velocity vector of the target to be tracked from the polar coordinate system to the rectangular coordinate system by xy orthogonal decomposition to obtain a velocity observed value, wherein the formula is as follows:
wherein r is the relative distance between the target and the radar, α is the azimuth angle between the target and the radar, v is the relative velocity between the target and the radar, subscript 1 represents the cycle number, then X1|1＝[x1|1，y1|1，vx1|1，vy1|1]T＝[x1,y1,vx1,vy1]T，P1|1Supplied by millimeter wave radar manufacturers;
if the cycle number n ≠ 1, it means that the result X calculated using Kalman filtering has been obtained in the previous cyclen|n，Pn|n；
Assuming that the target approximately moves linearly at a constant speed in one period, the predicted state vector of the target in the (n +1) th period can be predicted by equation (7):
namely, it is
Wherein T is a radar scanning period, xn+1|n、yn+1|n、vxn+1|n、vyn+1|nThe predicted state vector of the target for the (n +1) th cycle calculated from the nth cycle
The covariance is calculated by equation (8):
Pn+1|n＝FPn|nFT+Qn (8)
wherein QnA Gaussian covariance matrix with a mean value of zero;
judging whether the two adjacent periods observe the same target, comparing the predicted state vector of the target with the state vector obtained by actual observation, and setting the actual state vector of the target observed in the (n +1) th period as: xn+1＝[xn+1,yn+1,vxn+1,vyn+1]TThen the comparison formula is as follows:
wherein, Δ x, Δ y, Δ vx，ΔvyIs an error margin set by man, if Xn+1And
if equation (9) is satisfied, the target observed in the (n +1) th cycle and the target observed in the nth cycle are considered to be the same target; if the target is the same target, performing the next step, otherwise, considering that the target is lost, namely the target is noise;
calculating a Kalman filtering result, wherein the Kalman filtering result of the (n +1) th period is as follows:
Pn+1|n+1＝Pn+1|n-KnPn+1|n (11)
Kn＝Pn+1|n(Pn+1|n+Rn+1),Rn+1＝P1|1 (12)
wherein, KnFor Kalman gain, Rn+1The covariance matrix observed in the (n +1) th period is R because the error of each observation of the same millimeter wave radar is approximately considered to be equaln+1＝P1|1；
Repeating the third step and the fifth step until the target disappears or the period ordinal number reaches N, if the target disappears before the period ordinal number reaches N, the target is considered as noise, and the tracking is abandoned; if the cycle number reaches N, i.e. N +1 equals N, then X is outputn+1|n+1。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110348509.5A CN113096397B (en) | 2021-03-31 | 2021-03-31 | Traffic jam analysis method based on millimeter wave radar and video detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110348509.5A CN113096397B (en) | 2021-03-31 | 2021-03-31 | Traffic jam analysis method based on millimeter wave radar and video detection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113096397A true CN113096397A (en) | 2021-07-09 |
| CN113096397B CN113096397B (en) | 2022-04-12 |
Family
ID=76671817
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110348509.5A Active CN113096397B (en) | 2021-03-31 | 2021-03-31 | Traffic jam analysis method based on millimeter wave radar and video detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113096397B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113450580A (en) * | 2021-08-19 | 2021-09-28 | 浙江安沿科技有限公司 | Radar for monitoring traffic flow |
| CN114038191A (en) * | 2021-11-05 | 2022-02-11 | 青岛海信网络科技股份有限公司 | Method and device for collecting traffic data |
| CN114863682A (en) * | 2022-05-09 | 2022-08-05 | 合肥雷能信息技术有限公司 | Urban traffic flow monitoring feedback system based on millimeter wave radar |
| CN114999161A (en) * | 2022-07-29 | 2022-09-02 | 河北博士林科技开发有限公司 | Be used for intelligent traffic jam edge management system |
| CN115171385A (en) * | 2022-07-06 | 2022-10-11 | 合肥雷能信息技术有限公司 | Traffic incident detection system based on millimeter wave radar and video linkage |
| CN115206089A (en) * | 2022-06-07 | 2022-10-18 | 西安电子科技大学 | Variable lane diversion system based on 4D millimeter wave radar |
| CN116189116A (en) * | 2023-04-24 | 2023-05-30 | 江西方兴科技股份有限公司 | Traffic state sensing method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| SE503641C2 (en) * | 1995-05-09 | 1996-07-22 | Foersvarets Forskningsanstalt | Phase controlled radar system for tracking |
| EP0908861A2 (en) * | 1997-09-16 | 1999-04-14 | Siemens Aktiengesellschaft | Method for transmitting traffic information |
| CN1897015A (en) * | 2006-05-18 | 2007-01-17 | 王海燕 | Method and system for inspecting and tracting vehicle based on machine vision |
| CN107274668A (en) * | 2016-04-06 | 2017-10-20 | 北京瑞拓电子技术发展有限公司 | A kind of congestion in road modeling method based on vehicle detection |
| CN110688922A (en) * | 2019-09-18 | 2020-01-14 | 苏州奥易克斯汽车电子有限公司 | Deep learning-based traffic jam detection system and detection method |
| CN111932884A (en) * | 2020-08-13 | 2020-11-13 | 中国第一汽车股份有限公司 | Traffic jam recognition method and device, vehicle and storage medium |
-
2021
- 2021-03-31 CN CN202110348509.5A patent/CN113096397B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| SE503641C2 (en) * | 1995-05-09 | 1996-07-22 | Foersvarets Forskningsanstalt | Phase controlled radar system for tracking |
| EP0908861A2 (en) * | 1997-09-16 | 1999-04-14 | Siemens Aktiengesellschaft | Method for transmitting traffic information |
| CN1897015A (en) * | 2006-05-18 | 2007-01-17 | 王海燕 | Method and system for inspecting and tracting vehicle based on machine vision |
| CN107274668A (en) * | 2016-04-06 | 2017-10-20 | 北京瑞拓电子技术发展有限公司 | A kind of congestion in road modeling method based on vehicle detection |
| CN110688922A (en) * | 2019-09-18 | 2020-01-14 | 苏州奥易克斯汽车电子有限公司 | Deep learning-based traffic jam detection system and detection method |
| CN111932884A (en) * | 2020-08-13 | 2020-11-13 | 中国第一汽车股份有限公司 | Traffic jam recognition method and device, vehicle and storage medium |
Non-Patent Citations (4)
| Title |
|---|
| ZHANG PING 等: "Adaptive background updating algorithm for traffic congestion detection based on Kalman filtering and inter-frame centroid distanc", 《2015 IEEE ADVANCED INFORMATION TECHNOLOGY, ELECTRONIC AND AUTOMATION CONTROL CONFERENCE (IAEAC)》 * |
| 于德新 等: "《交通工程学》", 31 January 2019, 北京理工大学出版社 * |
| 李茂华等: "基于视频检测技术的城市交通拥堵判别模型", 《公路交通技术》 * |
| 王川童: "基于视频处理的城市道路交通拥堵判别技术研究", 《中国优秀博硕士学位论文全文数据库(硕士) 信息科技辑》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113450580A (en) * | 2021-08-19 | 2021-09-28 | 浙江安沿科技有限公司 | Radar for monitoring traffic flow |
| CN114038191A (en) * | 2021-11-05 | 2022-02-11 | 青岛海信网络科技股份有限公司 | Method and device for collecting traffic data |
| CN114863682A (en) * | 2022-05-09 | 2022-08-05 | 合肥雷能信息技术有限公司 | Urban traffic flow monitoring feedback system based on millimeter wave radar |
| CN115206089A (en) * | 2022-06-07 | 2022-10-18 | 西安电子科技大学 | Variable lane diversion system based on 4D millimeter wave radar |
| CN115171385A (en) * | 2022-07-06 | 2022-10-11 | 合肥雷能信息技术有限公司 | Traffic incident detection system based on millimeter wave radar and video linkage |
| CN115171385B (en) * | 2022-07-06 | 2022-12-27 | 合肥雷能信息技术有限公司 | Traffic incident detection system based on millimeter wave radar and video linkage |
| CN114999161A (en) * | 2022-07-29 | 2022-09-02 | 河北博士林科技开发有限公司 | Be used for intelligent traffic jam edge management system |
| CN116189116A (en) * | 2023-04-24 | 2023-05-30 | 江西方兴科技股份有限公司 | Traffic state sensing method and system |
| CN116189116B (en) * | 2023-04-24 | 2024-02-23 | 江西方兴科技股份有限公司 | Traffic state sensing method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113096397B (en) | 2022-04-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113096397B (en) | Traffic jam analysis method based on millimeter wave radar and video detection | |
| CN103176185B (en) | Method and system for detecting road barrier | |
| CN102881022B (en) | Concealed-target tracking method based on on-line learning | |
| CN104200657B (en) | A kind of traffic flow parameter acquisition method based on video and sensor | |
| US10451720B2 (en) | Object detection apparatus and object detection method | |
| CN101576952B (en) | Method and device for detecting static targets | |
| CN110542885B (en) | Millimeter wave radar target tracking method in complex traffic environment | |
| CN103593678B (en) | A kind of long-span bridge vehicle dynamic load distribution detection method | |
| CN104050818B (en) | The moving vehicle speed-measuring method of based target tracking and Feature Points Matching | |
| CN104574439A (en) | Kalman filtering and TLD (tracking-learning-detection) algorithm integrated target tracking method | |
| CN105744232A (en) | Method for preventing power transmission line from being externally broken through video based on behaviour analysis technology | |
| CN111199647B (en) | Monitoring video detection method for continuous lane changing and illegal turning of road vehicles | |
| CN113093191B (en) | Road vehicle detection system based on millimeter wave radar | |
| CN112749210B (en) | Vehicle collision recognition method and system based on deep learning | |
| CN105809954A (en) | Traffic event detection method and system | |
| CN109343051A (en) | A kind of multi-Sensor Information Fusion Approach driven for advanced auxiliary | |
| CN111562570A (en) | Vehicle sensing method for automatic driving based on millimeter wave radar | |
| CN1606033A (en) | Weak target detecting and tracking method in infrared image sequence | |
| CN113313012B (en) | Dangerous driving behavior identification method based on convolution generation countermeasure network | |
| CN117115752A (en) | Expressway video monitoring method and system | |
| CN104537691B (en) | Based on piecemeal in the same way speed add up optical flow field segmentation moving target detecting method | |
| CN113253262B (en) | One-dimensional range profile recording-based background contrast target detection method | |
| CN116110230A (en) | Vehicle lane crossing line identification method and system based on vehicle-mounted camera | |
| CN202904792U (en) | Intelligent visualized alarm system | |
| CN113128847A (en) | Entrance ramp real-time risk early warning system and method based on laser radar |
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 |