DE102011101990B3 - Method for determining relative drawbar angle in articulated train, involves coupling tensile unit and trailer by rotatably mounted shaft and two-dimensional image is captured by unit for imaging two-dimensional image - Google Patents

Abstract

The method involves coupling a tensile unit (4) and a trailer (5) by a rotatably mounted shaft. The two-dimensional image is captured by a unit (2) for imaging a two-dimensional image. The intensity along a trajectory is determined, where a point in time, in another image intensity along the trajectory is determined. The similarity between the intensity along trajectory is determined in the image and multiple shifted intensities and the relative angle is determined depending on drawbar of the similarity measure. An independent claim is included for an apparatus for determining the relative drawbar angle in an articulated train with a pulling unit.

Description

The invention relates to a method and a device for determining a relative drawbar angle in a articulated train according to the preamble of claim 1 and claim 7.

Link trains consist of at least one pulling part, for example a motor vehicle, and at least one trailer. The tensile part may be an active or passive tensile part, wherein an active tensile part has a drive source for generating a drive energy, for example an internal combustion engine. A passive pulling part can be, for example, a trailer which is in turn attached to another pulling part and to which a trailer is attached.

The tension member and the trailer may be mechanically coupled via a rotatably mounted drawbar, wherein the trailer can perform a rotational movement about a center of rotation, such as a coupling head at the end of a trailer hitch. The trailer hitch is usually fixedly arranged on the tension member, usually on or parallel to a central longitudinal axis of the tension member.

Guiding the articulated train requires due to the rotatability of the trailer additional attention of the articulated train driver. In particular, when reversing it can easily come to unwanted movements of the trailer, which can not be easily or intuitively balanced by the link train driver.

It is therefore desirable to simplify guidance of an articulated train by, for example, automatically intervening in a steering of the pulling part and / or outputting visual and / or optical and / or haptic guidance information.

A central variable for such assistance procedures is a drawbar angle of the articulated train, which describes an angle between a central longitudinal axis of the pulling part and a central longitudinal axis of the drawbar, which usually also corresponds to a central longitudinal axis of the trailer.

The DE 10 2005 042 957 A1 discloses that in a method for determining at least one drawbar angle and one trailer angle of a trailer of an articulated train with the aid of a laser scanner, the distances to measuring points on the front of the trailer are determined. Furthermore, a straight line of equalization is determined by all measuring points and a vertical line intersecting the center of the straight line is determined. With the help of the drawbar length and the position of the vertical, the pivot point of the drawbar is determined and determines an intersection of the vertical with an X-axis. Further, the angles are determined in the triangle formed from the origin, fulcrum and intersection.

The DE 10 2009 012 253 A1 discloses a method of assisting in reversing a towing vehicle and trailer. A driver assistance device disclosed in the publication comprises a means for detecting a trailer angle between the longitudinal axles of towing vehicle and trailer, which is not described in greater detail.

From the DE 10 2004 022 113 A1 a system for monitoring of motor vehicle trailers in driving is known, with a camera device for imaging the relative orientation of the rear end of the tractor and trailer front area, a Auswertesubsystem for the acquired data and interfaces for driving operation relevant processing of the evaluated data, wherein in the Auswertesubsystem one or more, in each case a specific relative orientation defining templates are stored, which serve the evaluation subsystem as reference data for associated orientation angles of the vehicle combination, wherein the Auswertesubsystem is configured by programming for performing correlations between images of each currently detected, relative orientation and the / the template (s) to determine information about the instantaneous orientation angle between the longitudinal and / or transverse axes and / or vertical axes of towing vehicle and trailer, and alignment angles, which on a indicate critical driving situation, to recognize, and to transmit appropriate signals to the interfaces for driving operation-relevant further processing.

From the DE 10 2008 045 436 A1 a system for determining a kink angle between a towing vehicle and a trailer is known, comprising an image capture device for capturing an image of the trailer, an image memory for storing a plurality of reference images of the trailer and a plurality of reference image kink angles, each reference image being associated with a reference image kink angle, and a processing unit coupled to the image capture device and the image memory, wherein the processing unit is configured to capture an image of the trailer using the image capture device, compare the captured image with the plurality of reference images of the image memory, and the bend angle depending on the comparison of the captured image with the reference images and the reference image kink angles.

From the post-published DE 10 2010 008 324 A1 is a combination of a motor vehicle and a non-automotive mobile electronic device with a processor and a camera known, the camera is directed from the motor vehicle on a trailer of the motor vehicle, the processor evaluates images taken by the camera of the trailer to a Determine positional relationship between the motor vehicle and the trailer.

The technical problem arises of providing a method and a device for determining a relative drawbar angle in an articulated train, which permit an image-based determination of the relative drawbar angle, wherein a determination of the relative drawbar angle which reduces the calculation time and calculates the relative drawbar angle while simultaneously ensuring high robustness of the determination is possible.

The solution of the technical problem results from the objects having the features of claims 1 and 7. Further advantageous embodiments of the invention will become apparent from the dependent claims.

A method for determining a relative drawbar angle in a train is proposed. The articulated train comprises a train part and a trailer. The tension member and the trailer are mechanically coupled via a rotatably mounted drawbar. The drawbar angle describes an angle between a central longitudinal axis of the towing vehicle or the tensile member and a central longitudinal axis of the trailer. Also, the drawbar angle can describe an angle between a central longitudinal axis of the pulling part and a central longitudinal axis of a drawbar section of the trailer. In this case, the trailer itself can be pivoted relative to the drawbar section, for example via a so-called turntable arrangement.

A device for image capture captures a two-dimensional image, wherein at least part of the drawbar is imaged in the image. Also, at least a part of the tensile part, preferably at least part of the trailer hitch, can be imaged in the image.

According to the invention, a first intensity and / or color gradient along at least one first trajectory is determined at a first time in a first image. The intensity profile can be, for example, a gray scale gradient. A color gradient can be, for example, a red-color profile and / or a green-content profile and / or a blue-content profile. The first trajectory forms a circular arc around the pivot point of the drawbar. The first trajectory may be, for example, partially circular or partially ellipsoidal in the image. The first trajectory can also be described by means of a mathematical function or via a stored number of pixels defining the trajectory. If the at least one device for image acquisition produces an exact plan view of at least part of the drawbar, then the first trajectory is part-circular, with a center being given by the pivot point of the drawbar, for example by the coupling head. If the at least one device for image acquisition produces a perspective view of at least part of the tensile part and at least part of the drawbar, which will generally be the case, the first trajectory may be distorted, for example partially elliptical, with respect to the circular course.

The drawbar and thus also the trailer will, especially on a flat roadway, move or pivot around the pivot point, with individual sections or points of the drawbar moving on part-circular trajectories. Depending on the position and orientation of the at least one image capture device such a trajectory is not part-circular, but z. B. partial ellipsoidal in the at least first image. A course of the first trajectory can therefore also be determined as a function of intrinsic and / or extrinsic imaging parameters of the at least one device for image acquisition.

The information of the intensity and / or color gradient along the first trajectory can z. B. stored in a memory vector with a predetermined number of memory locations.

Analogously, at least one further time in a further image, a further intensity and / or color gradient along the first trajectory is determined. Information, for example intensity values and / or color values, can be stored in a further memory vector.

Furthermore, a similarity measure between the first intensity and / or color gradient along the at least first trajectory in the first image and a number of shifted further intensity and / or color gradients along the at least first trajectory in the further image is determined. The shifted further intensity and / or color gradients are hereby shifted along the at least first trajectory in the further image. The displacement forms a circular motion with a predetermined angular extent along the circular arc imaged by the first trajectory. The number of shifted further intensity and / or color gradients can in this case cover a predetermined displacement range, which in turn is at a predetermined angular range along the By means of the first trajectory depicted circular arc corresponds, for example, to an angular range of -1 ° to +1. In this case, displacements with predetermined angular steps, for example 0.1 °, can be determined, which then form the number of shifted further intensity and / or color gradients. The number of shifted further intensity and / or color gradients can also comprise a further intensity and / or color progression which is not shifted, wherein an angle of the displacement amounts to 0 °.

A shift can take place, for example, by a shift of the values stored in the memory locations of the second memory vector with respect to these memory locations. For example, a shift of the values stored in the memory locations by one memory location may correspond to a predetermined angular step, for example 0.1 °.

A similarity measure between the intensity and / or color gradients can be determined, for example, by means of a correlation between the intensity and / or color gradients, for example between the storage vectors. As a result, a correlation coefficient can be determined as a similarity time. The similarity measure can also be determined by means of a normalized correlation between the first intensity and / or color progression and the number of shifted further intensity and / or color gradients.

The drawbar angle is then determined as a function of the individual similarity measures, which represent the similarity of the first intensity and / or color progression to the individual shifted further intensity and / or color progressions. For example, the drawbar angle can be determined as the angle corresponding to the displacement of the further intensity and / or color curve, which has a maximum correlation coefficient. Thus, advantageously, a relative change in the drawbar angle, ie a relative drawbar angle, can be determined as a function of the similarity measures.

The device for image acquisition can be fixedly arranged on the pulling part and have a previously known position and / or orientation with respect to a central longitudinal axis of the pulling part. By way of example, the device for image acquisition can be a reversing camera.

The inventive method results in an advantageous manner that a relative drawbar angle can be determined by a one-dimensional image analysis, namely an analysis of the intensity and / or color gradients along a trajectory. As a result, a computational effort in the determination of the intensity and / or color gradients as well as in determining the or similarity measures can be minimized. At the same time results in. Advantageously, a high degree of robustness. Overall, the inventive method can be performed in real time and with a low latency. Thus, high repetition rates of, for example, 30 Hz can be achieved with comparatively little hardware usage.

In a further embodiment, the further intensity and / or color gradient is shifted along the at least first trajectory in the further image using intensity and / or color interpolation. The intensity and / or color interpolation determines an intensity and / or color value of a pixel on the at least first trajectory as a function of intensity and / or color values of adjacent pixels, wherein the adjacent pixels lie in an interpolation range which is a predetermined amount in Radial direction to the first trajectory and extends by a predetermined amount along the first trajectory around the pixel in the other image. As a result, not only adjacent pixels along the first trajectory are taken into account for the purpose of interpolation, but also pixels which are not located exactly on the first trajectory but radially adjacent to the first trajectory. This advantageously results in an avoidance of aliasing effects of a pixel pattern and an improvement in the determination of the further shifted intensity and / or color gradients and thus also an improvement in the accuracy of the determination of the drawbar angle. In particular, by means of the interpolation, an angle determination can also take place in a subpixel area. In this way, in particular very small angular steps can be taken into account when shifting the further intensity and / or color gradients.

In a further embodiment, in each case a contrast and / or a brightness of the image are normalized before the determination or evaluation of the at least one intensity or color gradient in the first and in the further image. This advantageously results in a reduction of a contrast and / or brightness dependence in the determination of the relative drawbar angle according to the invention. For example, a contrast dependence can be reduced by locally z. B. using wavelet functions or similar functions, a texture energy is determined and used for local normalization. Also, methods that can determine an evaluation of a contrast as a function of a variance of an image signal can be used.

Alternatively or cumulatively, a normalization of a brightness z. B. be achieved by averaging the image values. For this purpose, the image signal z. B. with an equal-share wavelet function, z. Garbage function, or the like.

In a further embodiment, intensity and / or color gradients along a plurality of different trajectories are determined at the first and the further time in the two-dimensional image. The drawbar angle can then be determined as a mean drawbar angle, which results from the drawbar angles of the individual corresponding trajectories. Corresponding trajectories here designate trajectories in the first and in the further image, each of which depicts the same circular arc around the pivot point of the drawbar.

The drawbar angle can also be determined by means of a weighted mean value formation, wherein the drawbar angles resulting from the individual corresponding trajectories are weighted by means of a weighting factor. For example, the weighting factor can be determined as a function of a correlation coefficient of a normalized correlation. In this case, the correlation coefficient of a normalized correlation reproduces an absolute value via a similarity of intensity and / or color gradients along corresponding trajectories and therefore permits a corresponding weighting of the drawbar angle resulting from the corresponding trajectories. If, for example, the correlation coefficient of the normalized correlation is low, it can be assumed that the intensity and / or color gradients along the corresponding trajectories are not very similar, this angle being weighted in the following averaging with a low weighting factor. Conversely, tiller angles resulting from highly similar trajectories (high correlation coefficient) can be heavily weighted in averaging.

In a further embodiment, intensity and / or color gradients along a plurality of mutually different trajectories are evaluated at the first and the further time in the two-dimensional image, and a plurality of drawbar angles resulting from the corresponding trajectories are determined. For each of these drawbar angles, a number of similar drawbar angles are determined which lie in a predetermined angular range around the drawbar angle. An average drawbar angle is determined, which results from the drawbar angle with the maximum number of similar drawbar angles and these similar drawbar angles. Here, too, a weighted averaging, as previously described, can be used for the calculation of the average drawbar angle. This advantageously results in an increase in robustness in the determination of the relative drawbar angle. In particular, hereby significantly false partial results, ie tiller angle can be removed so that they have no negative impact on the overall result.

In a further embodiment, the displacement of the at least one further intensity and / or color gradient takes place along the at least first trajectory in the further image as a function of an absolute drawbar angle, the absolute drawbar angle being estimated using a motion model and / or determined from image information. The movement model can be designed as an observer. By means of the movement model, an absolute tiller angle can be estimated as a function of further measured and / or trailer-related measured variables, for example as a function of a yaw rate, a steering angle, a speed and other variables. From the difference of an estimated absolute drawbar angle at the first time point and the further time point, a starting value for a displacement of the at least one further intensity and / or color gradient along the at least first trajectory in the further picture thus advantageously results.

Similarly, an absolute drawbar angle can also be determined from image information, wherein z. B. the image generated by the device for image acquisition is evaluated with respect to a determination of the absolute drawbar angle. As a result, the method according to the invention can advantageously be further accelerated.

Overall, the inventive method for determining a relative drawbar angle can be determined without a so-called teach-in phase. Here, the method of the invention makes use of the property that the trailer and the drawbar from the perspective of the tension member concentric to z. B. rotate a coupling head (pivot point). Such behavior is then observed in the image data. Thus, when viewing a color and / or intensity distribution along concentric orbits around the coupling head, patterns of drawbar and road can be seen as a function of angle. Here, the pattern of the drawbar is the only structure that moves along the circular path. Structures of the road move, for example, on a path cutting the circular path. Thus, only the drawbar generates, apart from changes in brightness, a constant pattern when the trailer is pivoted.

However, a so-called teach-in phase or inclusion of data of the tensile component (steering angle, odometry or feedback of the control) can contribute to the improvement of an absolute accuracy.

The method according to the invention can be used for plausibility checking and error detection of further analysis results. For example, a check of an absolute tiller angle estimated by means of a movement model can be verified.

Advantageously, the relative drawbar angle determined by means of the method according to the invention can be used as an input variable for a device for controlling a driver assistance system, which, as described at the beginning, supports a linkage train.

Further proposed is a device for determining a relative drawbar angle. The device comprises at least one device for image acquisition and at least one evaluation device. By means of the proposed device, one of the previously described methods can be carried out in an advantageous manner. In particular, the device for image acquisition can be a reversing camera.

The invention will be explained in more detail with reference to an embodiment. The figures show:

1 a schematic block diagram of a device according to the invention,

2 a schematic representation of a drawbar angle,

3 a schematic representation of an articulated train and

4 a schematic representation of an inventive interpolation.

Hereinafter, like reference numerals designate elements having the same or similar technical characteristics.

In 1 is a device according to the invention 1 shown. The device 1 includes a device 2 for image capture, for example, a reversing camera. This is data technology with an evaluation 3 connected, for example, in a tensile part 4 (please refer 2 ) can be arranged.

In 2 an articulated train is shown, which consists of a train part 4 and a trailer 5 consists. The tensile part 4 here has a trailer hitch 6 with a coupling head 7 on. The trailer 5 has a drawbar 8th on, which is rotatable on the coupling head 7 is stored. Also shown is a central longitudinal axis 9 of the train part 4 and a central longitudinal axis 10 of the trailer 5 , A drawbar angle α describes an angle between the central longitudinal axis 9 and the central longitudinal axis 10 , Furthermore, there is a detection area 11 the device 2 shown for image capture.

In 3 a side view of an articulated train is shown. The tensile part 4 is here as a motor vehicle with a trailer hitch 6 , which in turn has a coupling head 7 has formed. In a rear area of the train part 4 is the device 2 arranged for image capture and schematically a detection area 11 this device 2 shown for image capture. The device 2 for image acquisition thus covers part of the trailer hitch 6 , in particular the coupling head 7 and part of the drawbar 8th of the trailer 5 ,

In 4 schematically an interpolation according to the invention is shown. Here is in 4 again a train part 4 and a trailer 5 represented by a coupling 6 with a coupling head 7 and a drawbar 8th mechanically coupled. A drawbar point DP moves with a pivoting of the trailer 5 around the coupling head 7 around along a circular arc 12 , To an intensity value of a pixel BP on this arc 12 can be interpolated between adjacent pixels P located in an interpolation region extending with a predetermined radial width Δr radial to the circular path 12 and with a predetermined length Δφ along the circular path 12 extends.

LIST OF REFERENCE NUMBERS

1

contraption

2

Device for image acquisition

3

evaluation

4

pulling part

5

pendant

6

clutch

7

coupling head

8th

shaft

9

central longitudinal axis

10

central longitudinal axis

11

detection range

12

arc

α

 drawbar angle

DP

pedestrian point

BP

pixel

P

neighboring pixels

.delta..sub.R

radial width

Δφ

length

Claims (8)

Method for determining a relative drawbar angle in an articulated train, wherein the articulated train is a tensile part ( 4 ) and a trailer ( 5 ), wherein the tensile part ( 4 ) and the trailer ( 5 ) via a rotatably mounted drawbar ( 8th ) are mechanically coupled, wherein a device ( 2 ) captures a two-dimensional image for image acquisition, wherein in the image at least a part of the drawbar ( 8th ), characterized in that at a first time in a first image, a first intensity and / or color gradient along at least a first trajectory is determined, wherein the first trajectory a circular arc ( 12 ) about a pivot point of the drawbar ( 8th ), wherein at least one further time in a further image, a further intensity and / or color gradient along the first trajectory is determined, wherein a similarity measure between the first intensity and / or color gradient along the at least first trajectory, in the first image and a number of shifted further intensity and / or color gradients is determined, wherein the shifted further intensity and / or color gradients are shifted along the at least first trajectory in the further image, wherein the relative drawbar angle is determined as a function of the similarity measures. A method according to claim 1, characterized in that a shift of the further intensity and / or color gradient along the at least first trajectory in the further image by using an intensity and / or color interpolation, wherein the intensity and / or color interpolation an intensity and / or color value of a pixel BP in dependence on intensity and / or color values of adjacent pixels P, wherein the adjacent pixels P are in an interpolation range, which is a predetermined amount Δr in the radial direction to the first trajectory and by a predetermined amount Δφ along the first Trajectory extends around the pixel BP in the other image. Method according to one of the preceding claims, characterized in that prior to evaluation of the at least one intensity and / or color gradient in the first and in the further image, a contrast and / or brightness of the image are normalized. Method according to one of the preceding claims, characterized in that at the first and the further time in the image intensity and / or color gradients along a plurality, different trajectories are determined, the drawbar angle as the average drawbar angle resulting from the drawbar angles, which result from the comparisons of the individual corresponding trajectories. Method according to one of the preceding claims, characterized in that evaluated at the first and the further time in the image intensity and / or color gradients along several different trajectories and several drawbar angles are determined, for each drawbar angle determines a number of similar drawbar angle which is in a predetermined angular range around the drawbar angle, determining a mean drawbar angle resulting from the drawbar angle with the maximum number of similar drawbar angles and those similar drawbar angles. Method according to one of the preceding claims, characterized in that the displacement of the at least one further intensity and / or color gradient takes place along the at least first trajectory in the second image as a function of an absolute drawbar angle, wherein the absolute drawbar angle estimated using a motion model and / or is determined from image information. Device for determining a relative drawbar angle in an articulated train, the articulated train being a tensile part ( 4 ) and a trailer ( 5 ), wherein the tensile part ( 4 ) and the trailer ( 5 ) via a rotatably mounted drawbar ( 8th ) are mechanically coupled, wherein the device ( 1 ) at least one institution ( 2 ) for image acquisition and at least one evaluation device ( 3 ), the device ( 2 ) captures a two-dimensional image for image acquisition, wherein in the image at least a part of the drawbar ( 8th ), characterized in that by means of the evaluation device ( 3 ), a first intensity and / or color gradient along at least one first trajectory can be determined at a first time, wherein the first trajectory is a circular arc ( 12 ) about a pivot point of the drawbar ( 8th ), wherein at least one further time in another image, a further intensity and / or color gradient along the first trajectory can be determined, wherein a similarity measure between the first intensity and / or color gradient along the at least first trajectory in the first image and a number of shifted further intensity and / or color gradients along the at least first trajectory in the further image can be determined, wherein the shifted further intensity and color gradients are shifted along the at least first trajectory in the further image, wherein the relative drawbar angle in dependence of the similarity measures is determined. Device according to claim 7, characterized in that the device ( 2 ) is a rear view camera for image capture.

Images