DE10239675B4 - Method for determining traffic state variables - Google Patents

Abstract

Method for determining traffic state variables, such as traffic flow velocity and traffic density, from gray values of a video image determined by a video camera, in which the video camera records video images of a traffic lane with vehicles, a measurement area being defined in the course of the imaged traffic lane in a video image and individual areas of the measurement area being determined Gray values are determined with which a gray value function G = f (g, s) is created, where g is the individual gray values and s is the distance of the respective individual gray value of a measuring origin, that below a transformation of the gray value function G in an equalized gray value function G _E , respectively the camera angle is carried out, so that a first gray value function G _E1 at a time t and one, separated by a temporal distance x, second gray value function G _E2 at a time t + x is determined, characterized in that each an arithmetic mean of e rst and the second gray value functions is determined and a first shift according to the respectively determined average values of the first and the second gray scale.

Description

The The invention relates to a method for determining traffic state variables, such as Traffic flow speed and traffic density, from gray values a video image determined with a video camera, in which the Video camera captures video images of a traffic lane with vehicles.

to solution the problem of automatic traffic condition identification, which when working on traffic management tasks it necessary to monitor traffic, traffic slumps to detect and to recognize the causes of this.

Present are especially systems in use that generate their information with induction loops. However, these only supply cross-sectional measured quantities. Farther Systems are known which use the Doppler effect a Determining the vehicle speed.

It is also common to Traffic observation video cameras use, for example Display pictures of current traffic on a monitor.

Known Image evaluation methods use various methods of object recognition, z. B. edge detection, pattern recognition and recognition of recognition License plates around statements about Generate traffic state variables.

The known solutions, As far as it is not about image processing, usually own to provide the disadvantage of no room-related sizes. In addition, let they do not make any conclusions on the cause of interference traffic. In the case of induction loops is Furthermore an intervention in the infrastructure by installation of the loops in the roadway required.

From the DE 41 28 312 A1 A method is known for detecting vehicle movements and traffic densities in traffic monitoring systems, in which a video camera used for traffic observation can be used in addition to counting and traffic data acquisition. The method is based on gray scale changes in successive video images with the disadvantage that non-traffic gray value changes occur in images, such as reflections on rainy roads and these influence the results of the process.

From the DE 196 40 938 A1 a traffic surface surveillance arrangement is described in which two video cameras are installed at a fixed distance from each other and aligned with the same overlapping observation field. With this arrangement, a more robust traffic measurement is possible, which is not affected by reflections or shadows on the traffic area.

In the JP 11328577 AA a method is described in which a surveillance camera captures a first image (a) and after a time difference Δt captures a second image (b). These images are digitized in an A / D conversion unit. For the area to be monitored, a brightness distribution is generated by the image recognition unit. Both brightness distributions include a caused by the camera position spatial Liche distortion, which is subsequently eliminated. These brightness distributions generated in this way can then be compared according to the method. For this purpose, the first brightness distribution is assigned to a second brightness distribution which corresponds to it and is displaced on a scale. After a successful assignment of the brightness distributions, the distance difference Δl on the scale is determined. Thus, in a subsequent step, the moving speed of the vehicles can be determined according to the known quantities .DELTA.t and .DELTA.l according to the relationship V = .DELTA.L / .DELTA.t.

adversely At this state of the art is that this comparison by means of a complex pattern comparison is realized.

In the DE 695 07 463 T2 an arrangement for determining traffic state variables is known with a video camera and a transmitting unit connected to the video camera, wherein the video camera is part of a local area network in which an evaluation unit and a transmitting unit are arranged as further network subscribers. In this case, the network subscribers are connected to one another via line-connected power lines and the transmitting unit is connected to an evaluation and / or display unit located outside the local network.

Of the Invention is therefore based on the object of specifying a method determined reliably with the traffic state variables which works self-calibrating and with which an automatic incident detection and information about their Causes possible is.

According to the invention, the object is achieved with a method for determining traffic state variables of the type described above by determining one arithmetic average of the first and the second gray value functions and performing a first shift in accordance with the respectively determined average values of the first and the second gray value function in that their respective arithmetic mean becomes zero, that one second displacement y is determined from a calculation of the maximum of a cross-correlation function of the first and second gray value functions and a determination of the abscissa distance of the maximum from the coordinate origin, which is the measure of the displacement y, that the calculation of the mean spatial traffic flow velocity V with V = y · M / x, where M is the scale in pixels per meter.

The method according to the invention uses the video image of a video camera installed above a traffic lane into which a measuring range determined in a suitable manner is projected. Single gray values g are generated from the measuring range consisting of parts. This can be, for example, the gray value of a pixel. These individual gray values have a fixed distance s to a measurement origin, which is preferably located at the bottom of the screen in the middle. The gray value function G is generated from these two variables. In a subsequent step, the gray value function is equalized, since the positioning of the video camera is not done perpendicular to the traffic lane but at a corresponding angle and thus distortion of the gray value function by spatial perception takes place. That is, the length of a vehicle at the bottom of the screen corresponds to a different number of pixels than the length of the same vehicle at the top of the screen. Accordingly, a, depending on the viewing angle of the video camera on the measuring range equalization of the gray value function G is performed and the equalized gray value function G _E generated. For a calculation of the mean spatial traffic flow velocity V, two of these equalized gray value functions are necessary, of which it is known at which time interval x they were generated from each other. The shift y (in image pixels) of the gray value functions to one another is determined by cross correlation, so that V can be calculated with the aid of the scale M (in pixels per meter).

In one embodiment of the invention, it is provided that a normalized gray value function G _{EN is} generated by centering the gray value function G _E around its arithmetic mean value, that the gray value function G _EN eliminates values by zeroing which are smaller than a reference value and that the determination of the gray value function G _EN Traffic density over the number of zero crossings of the eliminated gray value function takes place. It has been found that a vehicle can be determined by two to three zero crossings.

To generate a gray value function G _EN with an arithmetic mean value of zero, this is determined and subsequently carried out a shift of the equalized gray value function G _E. This shift is a prerequisite for the determination of the traffic density, which takes place via the number of zero crossings of the gray value function. Another requirement is met by the removal of noise components from the gray value function. For this purpose, all signal components which fall below a reference value adapted to the instantaneous maximum value are blanked out by zeroing.

In A further embodiment of the invention provides that the measuring range is preferably in the lane center.

The Location of the measurement area in the center of the lane increases the probability that each vehicle moves within the measuring range and thus can be detected by the procedure, even if it is assumed is that not every vehicle is guided exactly in the lane center.

In an embodiment of the invention is provided that the Measuring range of punctiform, a measuring line forming, sub-areas and the gray scale correspond to the pixels lying on the measurement line.

In In this case, the measuring range is one of several pixel-shaped subregions existing measuring line, which can be a straight line. In an uneven Course of the lane, for example in a curve, an adjustment the measuring line to the lane course to ensure that the measuring line runs in the lane center.

In an execution the invention provides that the measuring range of several in longitudinal direction successive sheet-like subregions exists and the gray values over the respective subregions correspond to averaged gray values.

One Subarea may consist of several side by side and / or one above the other lying pixels exist. These, forming a surface sections are in longitudinal direction arranged the lane. The arrangement will depend on the Course of the lane, also in this case, preferably in the lane center respectively.

In a cheap one execution The invention provides that several lanes of a video image be evaluated simultaneously and that for each measuring range one each Gray value function G is generated.

The tracking and evaluation of the traffic state variables is not limited to a measuring range or a lane. Within a video image, several measuring ranges can be defined be laid, the course of which is not restricted. For example, the video image of an intersection may include a nearly horizontal and a vertical measurement range.

In a further embodiment of the invention it is provided that the reference value is one third of the maximum value of the gray value function G _EN .

In A particular embodiment of the invention is provided that the reference value is adjusted periodically.

Of the Maximum value of the gray value function is a measure of the current gray value of the Vehicles. Each gray value of less than, for example, a third of the maximum value no information about Vehicles and is eliminated by zeroing. By a periodic Determination of the reference value, the adjustment of the procedure changing Conditions (lighting conditions).

In An embodiment of the invention is provided that for the determination the traffic density is determined by the number of vehicles the number of zero crossings the gray value function is divided by 2.5.

By Measurements have been made that a vehicle is driven by a vibration with two to three zero crossings is mapped in the gray value function. This is the number of Zero crossings a measure of the number of the vehicles in the measuring range. The number itself is obtained by Division of the number of zero crossings by 2.5. The use of the Quotient 2.5 taken into account, that a vehicle is represented by two to three zero crossings.

In a particular embodiment of the invention is provided that an error signal is emitted when vehicles are detected on the lane when the traffic flow rate is below a threshold decreases.

Sinks the traffic flow rate determined by the inventive method, For example, on a highway, below a certain threshold This is assumed to be 40 km / h in this example Procedure an output noise ready, which is an input signal for an automatic traffic control device or with manual monitoring an optically / acoustically operating signal module can be, after whose occurrence the permissible top speed is lowered on the lane.

A Arrangement for determining traffic state variables may be designed such that the video camera is part of a local network in which one Evaluation unit and a transmitting unit as another network participants are arranged, the network participants via wired power lines connected to each other, and the transmitting unit with an outside the local network lying evaluation or display unit is connected.

A Such an arrangement includes, for example, a video camera for Detecting the traffic area, one immediately near the video camera arranged evaluation and a transmitting unit. The transmitting unit transmits information to an evaluation and / or display unit. The transmission takes place by cable or by radio. The evaluation and / or display unit is with a Number of such assemblies connected. Through this unit can then be controlled, for example, the display of traffic areas will be a disruption was detected. The arrangement of the evaluation unit directly on the video camera allows the recording of traffic in small time units of for example one second. Through the evaluation takes place the procedural processing the video data without caching the images. In the result the processing is then traffic status data, Störungsmeldun conditions, a "level of service "or Video images available, transmitted by the transmitting unit at intervals of one minute to the Evaluation and / or display unit to be sent. Through the "level of service " for example, the states "liquid traffic", viscous traffic or "traffic jam" classified.

The Connection between transmitting unit and evaluation and / or display unit Can be conducted by wire.

In a form of tight spatial Arrangement of the assemblies, for example by the evaluation and Transmitter unit are combined in an assembly, processing is the data in real-time operation allows also with several lanes to be monitored simultaneously in a video picture.

The Invention will be explained in more detail with reference to an embodiment. In the associated Drawings shows

1 a traffic area with vehicles and a measurement area formed by a straight measuring line,

2 a traffic area with vehicles and a measuring area formed by several areas, shifted to the vertical perpendicular,

3a a top view of a traffic lane,

3b an associated gray value function along the measurement line,

4a a gray value function with its arithmetic mean,

4b a centered gray value function,

5 a representation of the camera point of view with distorted and equalized gray value function,

6a a gray scale function at time t,

6b a gray value function at the time t + x

7 a result of a cross-correlation of two gray value functions,

8a a mean-free, equalized gray value function with a reference value,

8b a gray value function with eliminated signal components and

9 an arrangement for implementing the method according to the invention.

One over a lane 1 installed video camera pointing to the middle of the lane 1 on which there are vehicles 6 are located, aligned and inclined to this at an angle of, for example, 45 degrees, takes pictures of the traffic happening. These are available in digital form or can be generated by a downstream module and on a control monitor 2 be reproduced. Within this video image, one or more measurement ranges are defined, each of which consists of subregions. These subareas are single, one measurement line 3 forming pixels, whereby also a measuring surface 4 can be considered as a subarea. The lowest point of this imaginary line of measurement 3 on the monitor 2 is the measurement origin 5 to which each pixel lying on the measurement line has a certain distance s, which also corresponds to a certain path on the lane. The gray value function G = f (g, s) 7 is by measuring the on the measuring line 3 lying individual gray values g, in the present example, the pixels produced. In a subsequent step, the gray value function is equalized 7 because the positioning of the video camera is not perpendicular to the lane 1 but at an angle of 45 degrees and thus a distortion of the gray value function 7 through spatial perception. That is, the length of a vehicle 6 at the bottom of the screen is a different number of pixels than the length of the same vehicle 6 at the top of the screen. As a result, one becomes, from the viewpoint of the video camera 10 Equalization of the gray value function G dependent on the measuring range 7 performed and the equalized gray value function G _E 11 generated. To calculate the mean spatial traffic flow velocity V, two gray value functions G _E1 generated at a known time interval x are calculated 14 and G _E2 15 needed. The displacement of the two gray-value functions relative to one another represents a distance traveled by the vehicles over time x 6 in the lane 1 This is determined by a cross-correlation function 18 from both gray value functions 14 and 15 is determined by the displacement of the Maximus 17 from the coordinate origin of the abscissa, the shift of the gray value functions y 16 in pixels. With the known scale M in pixels per meter, which is the ratio of the distance between two adjacent pixels on the measurement line 3 to an associated path on the pictured lane 1 represents, after V = y · M / x, the traffic flow velocity is determined.

If it is determined after the determination of the speed that it is below a limit, which is assumed at 40 km / h on a highway, the generation of a fault signal. This can be processed by a traffic control system and the maximum permissible speed limit displays on the affected lane (s) 1 or by the output of an acoustic and / or optical signal, during manual traffic monitoring in a central office, the present camera image 2 evaluated and corresponding actions initiated.

at a central traffic monitoring can not all images of the system video cameras be played back on monitors. In this case, this allows inventive method to control the playback. Occurs in an area a traffic intrusion on, so does the associated Video image automatically displayed.

Disruptions in the determination of the traffic state variables, such as shadows or reflections by moisture, which can simulate vehicles, do not disturb the inventive method, since only at a minimum speed of for example 10 km / h is assumed that vehicles 6 on the lane 1 are located.

For a determination of the traffic density, the equalized gray value function G _E 11 around its arithmetic mean 8th centered. This normalized (mean-free) gray value function G _EN 12 contains information about vehicles 6 and disturbances characterized by a much lower amplitude. Since the determination of traffic density over the number of zero crossings of the gray value function G _EN 12 takes place, these noise components must be eliminated. This is the current maximum value 21 the gray value function which is a measure of the current gray value of the vehicles. Starting from the maximum, becomes a reference value 20 that is at one third of the maximum 21 lies, determined. All values of the gray value function G _EN 12 that this reference 20 fall below are eliminated by zeroing. This process produces a grayscale signal 19 in that essentially only contains components in the signal generated by vehicles 6 were caused. The maximum 21 and thus the reference value 20 is determined periodically and thus adapts to changing conditions (lighting conditions).

In investigations carried out it was determined that a vehicle 6 is imaged by a vibration with two to three zero crossings. With this knowledge can the number of vehicles 6 be determined by the number of zero crossings of the gray value function 19 divided by 2.5.

1

lane

2

control monitor

3

measuring line

4

measuring surface

5

measurement origin

6

vehicles

7

Gray value function G

8th

arithmetical Average

9

centered Gray value function

10

perspective the camera

11

equalized gray value function G _E

12

normalized and equalized gray value function G _EN

13

camera image

14

Gray value function G _E , at time t (in s)

15

Gray value function G _E2 at time t + x (in s)

16

y shift of the gray value function G _E , at the time t (in s) compared to the gray value function G _E2 at the time t + x (in s) in image pixels

17

maximum value the cross-correlation function between the shifted by y pixels pixel gray scale functions

18

Cross-correlation function between the shifted by y pixels pixel gray value functions G _E1 and G _E2

19

Gray value function with eliminated non-relevant signal components

20

reference value for eliminating non-relevant signal components

21

Maximum value of the gray value function G _EN

22

video camera

23

local network

24

evaluation

25

transmission unit

26

Conducted power lines

27

evaluation and / or display unit

Claims (10)

Method for determining traffic state variables, such as traffic flow velocity and traffic density, from gray values of a video image determined by a video camera, in which the video camera records video images of a traffic lane with vehicles, a measurement area being defined in the course of the imaged traffic lane in a video image and individual areas of the measurement area being determined gray values are determined with which a gray value function g = f (g, s) is created, where g is the single gray-scale values and s is the distance of each individual gray value of a measurement origin, that subsequently a transformation of the gray value function g in an equalized gray value function g _e, corresponding to the camera angle is carried out, so that a first gray value function G _E1 at a time t and one, separated by a temporal distance x, second gray value function G _E2 at a time t + x is determined, characterized in that each an arithmetic mean of first and second gray value functions are determined and a first shift takes place in accordance with the respectively determined mean values of the first and the second gray value functions such that their respective arithmetic mean value becomes zero, that a second shift y is calculated from the calculation of the maximum of a cross correlation function (FIG. 17 ) of the first and second gray value functions ( 14 . 15 ) and a determination of the abscissa distance of the maximum ( 17 ) from the coordinate origin, which determines the measure for the displacement y ( 16 ), it is determined that the average spatial traffic flow velocity V is calculated as V = y * M / x, where M is the scale in pixels per meter. Method according to Claim 1, characterized in that a normalized gray value function G _EN ( 12 ) by centering the gray value function G _E ( 11 ) by their arithmetic mean ( 8th ) is generated by the gray value function G _EN ( 12 ) Values are eliminated by zeroing which is less than a reference value ( 20 ) and that the determination of the traffic density over the number of zero crossings of the eliminated gray value function ( 19 ) he follows. Method according to claim 1, characterized in that that the measuring range is preferably in the lane center. A method according to claim 1, characterized in that the measuring range of punctiform, a measuring line ( 3 ), and the individual gray values on the measuring line ( 3 ) lie corresponding pixels. A method according to claim 1, characterized in that the measuring range of a plurality of longitudinally consecutive sheet-like portions ( 4 ) and the individual gray values over the respective subregions ( 4 ) correspond to averaged gray values. Method according to claim 1, characterized in that several lanes ( 1 ) of a video image ( 13 ) are evaluated simultaneously and that for each measuring range ( 3 or 4 ) one gray value function G ( 7 ) is produced. Method according to claim 2, characterized in that the reference value ( 20 ) one third of the maximum value of the gray value function G _EN ( 21 ). Method according to claim 2, characterized in that the reference value ( 20 ) is adjusted periodically. A method according to claim 2, characterized in that for determining the traffic density, the number of vehicles ( 6 ) is determined by the number of zero crossings of the gray value function ( 19 ) is divided by 2.5. A method according to claim 1, characterized in that in a detection of vehicles ( 6 ) on the lane ( 1 ) an error signal is output when the traffic flow velocity V falls below a threshold.