DE102011053284B3 - Method for verifying speed of appropriate vehicle, involves calculating image angle based on displacement of vehicle in images taken at preset subsequent times and intrinsic parameters of camera, and comparing with object angle - Google Patents

Abstract

The invention relates to a method in which the detected speed (v) of a vehicle is verified in that an object angle (φO) which a vehicle driving through the object area of the camera includes with a camera (K) offset in time between at least two points in time, and an image angle (φO), which the images of the vehicle include between consecutive recordings of the camera (K) at these at least two points in time, is calculated and the object angle (φO) and the image angle (φO) are compared with one another.

Description

In traffic monitoring, invasive and non-invasive measuring sensors (eg induction loops, piezo strip laser sensors, radar sensors) are used to measure the speed of vehicles. The measured values thus obtained can be speed values or distance values or time values from which the speed of the appropriate vehicle can be derived. Speed detection can be used for statistical purposes, but is largely used to detect and punish traffic violations. It is customary to document the traffic violation by the image of the infringement vehicle is created, in which usually legally relevant data, such as the detected speed, the date and time are displayed.

The invention relates to a method with which, in particular, the correctness of a speed detected by a vehicle can be confirmed or a vehicle which is imaged in an image recording and moves at the detected speed can be identified. In this case, a secure identification is given only if the speed can be assigned to a vehicle mapped one-to-one.

A one-to-one correspondence means that the detected speed can be assigned exactly to a mapped vehicle and also only this vehicle to the detected speed. Such a one-to-one assignment is not only required when multiple vehicles are shown in the image capture, but also, if only a single vehicle is shown, as shown here by measuring errors z. B. by jet reflection, a speed can be measured, which does not correspond to the actual speed of the imaged vehicle.

In order to realize the explained assignment, there are various methods known in the art.

In the case of speed measurement, both with non-invasive sensors, whose measurements are used not only to record the speed but also the distance of a vehicle, as well as to invasive sensors, such as sensors. B. two piezo strips whose distance from each other and distance to a camera, with which the images are created, is known, an assignment of the detected speed over the distance can be done. The distance is the distance to the sensor, and due to the knowledge of the relative position of the sensor to the camera, the distance to the camera can be derived. In the following, because it is immaterial to the description of the invention, it should be assumed in simplified form that the sensor or, in the case of a plurality of sensors arranged for speed detection, eg. B. two piezo strips, which is positioned in the direction of travel last sensor, in the same place as the camera. This can be understood below by distance, both the distance of the vehicle to the measuring sensor and the camera.

For non-invasive sensors, where the distance of a vehicle is detected permanently while passing through the measuring range, the distance detected at the time the camera is triggered can be used for identification.

For invasive sensors, from the known distance of the sensor and the detected speed, the probable distance of the vehicle at the time the camera is triggered can be determined.

With knowledge of the intrinsic parameters of the camera as well as the position and orientation of the camera axis of the camera to the roadway and the sensor, the detected distance can be assigned to a vehicle depicted in the image recording. However, this is not uniquely possible in particular if the image recording includes several lanes of the roadway, so that at the same time several vehicles can be located at the same distance.

This can happen even if, by means of kinescoated reflection, another vehicle than the vehicle depicted has become adequate, so that a different distance and, therefrom, a different speed is derived, which was not caused by the imaged vehicle.

To exclude such uncertainties are z. B. Methods for non-invasive sensors are known in which detected with a measuring radiation, the speed and distance of a vehicle and the appropriate vehicle is tracked, d. H. The changing position data of the vehicle during passage through a measuring range are detected several times. The tracking data thus obtained describe the traffic lane (movement line) which the vehicle describes while passing through the measuring range, with which the tracking data can be used to identify the appropriate vehicle.

Usually z. B. on the knowledge of the lane, the vehicle can be assigned to a specific lane and thus identified in a picture image of a traffic scene. This assumes, however, that the roadway in the Image capture is divided into several lanes visible.

From the DE 10 2007 022 373 A1 A method for detecting traffic violations by detecting object tracking data by means of a radar sensor is known, which directs radar radiation onto a roadway in such a way that multiple vehicles can travel through a measuring area defined by the radar beam (radar lobe).

For this purpose, a radar device can be positioned next to the roadway or above the road, z. B. be attached to a bridge, in which case the horizontal axis of the radar axis coincides with the road direction.

From the reflected radar signals, the relative velocity of an object to the radar sensor, the distance of its radar radiation reflecting surfaces to the radar sensor and the angle at which the reflected radar radiation to the radar axis incident on the radar sensor, derive.

In a Dauerradarsensor of which in the DE 10 2007 022 373 A1 is assumed, there is a continuous speed measurement or a continuous distance measurement by utilizing the Doppler radar effect or the Frequenzumtastungsprinzips (FSK) in evaluation of the phase difference of reflected radar signals of different frequency. An angle measurement takes place z. B. by means of two receiving antennas via a triangulation measurement.

If during the measurement a speed is detected which is above a predetermined limit speed, the computer determines a lane from the detected distance and angle values and sends a signal to a camera for taking a picture of the current traffic scene. The camera is positioned and adjusted at a known fixed distance from the radar so that the optical axis (hereinafter camera axis) is aligned in fixed angular relation to the radar axis and the traffic scene is sharply imaged over a range of depths of focus by a predetermined distance, referred to as the photopoint ,

Since the object field of the camera extends over all lanes, over which also the radar cone is directed, also several vehicles can be imaged in the record, which are at the time of triggering the camera in the radar cone.

In order to unambiguously identify the infringing vehicle in the recording, the lane determined by the infringing vehicle is superimposed on the recording. The insertion advantageously takes place in such a way that a marking representing the lane is superimposed on the picture elements in the recording, which are to be assigned to the positions defined over the distances and the angles, which together form the lane, in the object field. That is, the identification is carried out solely on the basis of the radar relative measured lane without absolute reference to a single lane of the road.

Advantageously, several proof photos should be created during the passage of an infringer vehicle, which then show this at different positions of the determined lane.

Assuming that by a method according to the DE 10 2007 022 373 A1 Due to shadowing of the vehicles with each other not every offense can be conclusively recorded, will be published in the later DE 10 2010 012 811 A1 in addition to an object tracking as in the DE 10 2007 022 373 A1 is described by means of a Videokamea with a matrix receiver performed an image tracking and checks the obtained tracking data for correlation. In the case of correlation, a partially overlapping marker is displayed in the images, on a pictured vehicle.

During image tracking, the tracking data for determining the lane of a vehicle from the change in position of an imaged vehicle are determined.

An image tracking according to the method of post-published DE 10 2010 012 811 A1 is comparatively compute and memory intensive.

In the published patent application WO 2010/043252 A1 A method for determining the speed of vehicles is described which records a first image when exceeding a maximum speed and a second image of the appropriate vehicle at a time interval following. The time interval is specified as a function of the measured speed. From the image information, a velocity value is calculated, which can be used to check the plausibility of the measurement.

Another method for determining the speed of vehicles is disclosed in the publication DE 22 11 462 A disclosed. By means of known methods, the speed of a vehicle is measured here and, in addition, its position is registered at two different times. The registration becomes on the one hand the Calculation of the speed and on the other hand used to identify the vehicle.

In the published patent application GB 2342800 A a method for determining the speed of vehicles is described in which two, separated by a fixed time interval images of the vehicle to be measured are taken together with the traveled lane section. The calculation of the speed of the vehicle is based on the two images, by the measurement of the vehicle position with respect to applied on the roadway and known markings.

The invention has for its object to find a method with a relatively low computational and memory costs.

This object is created for a method for verifying the detected speed of a vehicle by means of a camera in which a camera whose location and orientation to a roadway on which vehicles pass through the object area of the camera, and thus to the direction of travel, several shots be solved in that at least at a time t ₁ , the distance of the vehicle e (t ₁ ) is determined to the camera to calculate depending on the detected speed v and in knowledge of the direction of travel, an object angle φ _O , the the vehicle has moved relative to the camera between time t ₁ and at least one later time t ₂ , and that a respective picture is taken with these at least two times t ₁ and t ₂ , and from the offset s' of the images of the vehicle A '(t ₁ ) and A' (t ₂ ) in the recordings and the intrinsic parameters of the camera an image angle φ _{B is} calculated, d it is compared with the object angle φ ₀ .

Depending on the measuring sensor used for the speed measurement, the distance measurement can also be carried out with this or another measuring sensor. The speed measurement can be done immediately before or at the same time as the verification procedure.

In the image of only one vehicle in the images and identity of object angle φ _O and image angle φ _B , within a predetermined tolerance range, it can be concluded that the imaged vehicle has actually moved the detected speed to z. B. to exclude measurement error by kink beam reflection.

When imaging multiple vehicles in the images, the detected speed can be assigned to the imaged vehicle, the image of which is shifted by an image angle φ _{B that} is identical within a predetermined tolerance range with the object angle φ ₀ .

The invention will be explained below with reference to an embodiment with reference to a drawing. This shows:

1 a graphic sketch for explaining the method.

With a method according to the invention, both a speed detected with any measuring sensor can be verified by detecting the distance of suitable vehicles with a further measuring sensor whose location is known in relation to the camera K, or by verifying a speed which coincides with a speed same measuring sensor is detected as the distance.

The detection of the speed v and the distance e to the camera K while passing through a surveillance area is, as explained in the introduction, possible with various invasive and non-invasive sensors. In order to be able to verify the determined speed v according to the invention or to be able to assign the speed v to a vehicle depicted in a photograph, in addition to the speed v, the distance e of the appropriate vehicle to a camera K has to be at a time t ₁ , as explained below is triggered, the camera K, determined and stored at least temporarily. The determination of the distance e is expertly determined depending on the measuring sensor used and the measuring principle.

From the knowledge of the location and the orientation of the camera K to the road and thus to the direction of travel, the distance e of the appropriate vehicle to the camera K at a time t ₁ , and the detected speed v, can be concluded on the traveled travel s, the Vehicle must have covered at a given later time t ₂ . In a manner known to those skilled in the art, the angular range, determined by the object angle φ _O , which the vehicle has crossed between the times t ₁ and t ₂ with respect to the camera K can thus be calculated.

The accuracy of the angle determination can be increased if also the distance e at time t _{2 is included} in the calculation. That is, it does not matter to determine the exact location of a vehicle at certain times, which is not always possible on the basis of the distance, but it is explained as an object angle φ _O determined. Only to the graphic representation from concrete locations where the vehicle is located.

In 1 is the location of a vehicle A (t ₁ ) at time t ₁ and A (t ₂ ) at time t ₂ and the respective associated distance e (t ₁ ) and e (t ₂ ), the traveled travel distance s = f (v) and the resulting object angle φ _{O is} shown.

Is used as a sensor for speed detection z. B. a radar sensor used, then from the knowledge of the location and the orientation of the radar sensor, determined by the radar angle of the main beam with the edge of the road includes, z. B. 20 ° and the knowledge of the location and orientation of the camera K, determined by the camera axis, relative to the radar sensor, and detected by the radar sensor distance and the distance of the vehicle to the camera K at the time of shooting derivable.

Be used as a sensor for speed detection z. B. used two piezo strips, then the distance of the vehicle to the camera K at the time of shooting from the knowledge of the distance between the second piezo strip and the camera K and the then known speed v at times of shots after passing the second piezo strip are derived ,

The camera K runs in parallel to the measured value acquisition. It can be a video camera or a single-frame camera that produces at least two temporally successive recordings. Preferably, it is started with the determination of a speed v, which is above a predetermined limit speed. The camera K must be aligned with the roadway so that its object field overlaps the surveillance area at least partially. The camera K stores within the period in which a vehicle is detected by the sensor, the monitoring area, at least two shots in which the appropriate vehicle is shown.

In 1 Further, the location of the imaged vehicle A '(t ₁ ) at time t ₁ , in a first shot and A' (t ₂ ) at time t ₂ in a second shot in the image plane BE of the camera K and the offset of the images of the vehicle s', which is dependent on the speed v and the distance e shown. From the offset s' and the intrinsic parameters of the camera K can be calculated in a manner known to those skilled in the angle range, determined by the image angle φ _B , by which the image of the vehicle is shifted in the images. The higher the resolution of the receiver matrix of the camera K, the more accurate the offset s' can be determined.

The calculated angle of view φ _B can then be compared with the calculated object angle φ ₀ , wherein a match within a predetermined tolerance confirms that the imaged vehicle from whose image shift the image angle φ _{B was} calculated has traveled at the measured speed v.

On the one hand, the correctness of the speed measurement can be confirmed for only one vehicle and, on the other hand, if several vehicles are depicted, the appropriate vehicle can be detected.

The offset s' is determined by calculating optical flow vectors between the images. For this purpose, methods of digital image processing are used which determine the offset s 'in pixels between at least two temporally successively recorded images of the appropriate vehicle (image offset s').

By knowing the intrinsic parameters of the camera K, it is also possible to calculate back to the offset s' of the vehicle (object offset).

The real offset of the vehicle calculated with knowledge of the angle of the camera's distance from the vehicle and the measured speed v can then be compared with the object offset calculated from the image offset s'. A correspondence confirms that the depicted vehicle is indeed the appropriate vehicle and that it is actually driving the measured speed v.

It is also possible to estimate the speed v of the vehicle from the quotient of the offset s' and the time duration between these images in order to be compared with the measured speed v for verification.

The intrinsic parameters of the camera can be measured before the start of the procedure, or they can be determined by measuring some vehicles and comparing them with the values derived from the images taken.

The time interval between two successive images should be small in order to keep the offset s' small. However, since only a small local offset s' tends to be noisy, it is preferred to create a sequence of recordings with more than just two recordings.

With a method according to the invention, the verification can take place with only two temporally successive recordings of the vehicle. In any case, in order to prove this in traffic safety, it is typical to have two images in close succession. In this respect, this method can be used for the already recorded image material and the use of an additional video system including the associated effort for the hardware (eg, evaluation computer and lighting) can be circumvented.

LIST OF REFERENCE NUMBERS

• A (t ₁ )

  Location of a vehicle at time t ₁

  A (t ₂ )

  Location of a vehicle at time t ₂

  e

  distance

  e (t ₁ )

  Distance at time t ₁

  e (t ₂ )

  Distance at time t ₂

  φ _O

  object angle

  s

  traveled route

  v

  speed

  A '(t ₁ )

  Location of the depicted vehicle at time t ₁

  A '(t ₂ )

  Location of the pictured vehicle at time t ₂

  K

  camera

  BE

  Image plane of the camera K

  s'

  Offset of the pictures of the vehicle

  φ _B

  angle of view

Claims (5)

Method for verifying the speed (u) of an adequate vehicle by means of a camera (K), with which a camera (K), its location and orientation to a roadway on which vehicles pass through the object area of the camera (K), and thus to Direction of travel, it is known, several shots are created, characterized in that at least at a time t _1, the distance of the vehicle (e (t ₁ )) to the camera (K) is determined to vary depending on the speed (v) and in Knowledge of the direction of travel to calculate an object angle (φ _O ) by which the vehicle between the time t ₁ and at least a later time t _{2 has moved} relative to the camera (K), that at these at least two times t ₁ and t ₂ with the camera (K) is in each case a recording is created, and from the offset (s') of the images of the vehicle A '(t ₁ ) and A' (t ₂ ) in the images and the intrinsic parameters of the camera (K) an angle of view (φ _B ) err is compared, which is compared with the object angle (φ _O ). A method according to claim 1, characterized in that the speed (v) is detected simultaneously with the implementation of the method for verification with a same measuring sensor, with which the distance is determined. A method according to claim 1, characterized in that the speed (v) is detected immediately before performing the method for verification with another measuring sensor, as with which the distance is determined. A method according to claim 1, characterized in that when imaging only one vehicle in the images and identity of object angle (φ _O ) and image angle (φ _B ) within a predetermined tolerance range it is concluded that the imaged vehicle actually the detected speed (v) drove to z. B. to exclude measurement error by kink beam reflection. A method according to claim 1, characterized in that when imaging multiple vehicles in the images, the detected speed (v) is assigned to the imaged vehicle whose image is shifted within a predetermined tolerance range with the object angle (φ _O ) identical image angle (φ _B ) is.

Images