DE102016220559A1 - Method and system for determining a roll angle of a two-wheeler during cornering - Google Patents

Abstract

The invention relates to a method and a system for determining a roll angle (λ) of a two-wheeler (1) during cornering. The method comprises the steps of recording an image sequence of an environment of the bicycle (1) by means of a camera (3) mounted on the bicycle (1), extracting an optical flow from the image sequence, determining the roll angle (λ) of the camera (3) with reference to FIG extracted optical flux and determining the roll angle (λ) of the two-wheeler (1) based on the determined roll angle (λ) of the camera (3).

Description

The invention relates to a method and a system for determining a roll angle of a bicycle during cornering.

From the EP 1 989 086 B1 a method and a device are known, which are used for determining a roll angle of a motorcycle, wherein a roll rate of the motorcycle is determined, in particular by means of a rotation rate sensor. From the roll rate, a first roll angle variable is determined, and at least one further dynamic driving parameter of the motorcycle is determined based on which at least one second roll angle variable is determined. From the roll angle variables, the roll angle is calculated, in particular by addition.

Furthermore, the teaches WO 2011/063785 A1 a method for estimating the roll angle in a moving vehicle. In a step a), a sequence of images of the vehicle surroundings is recorded with a camera, in particular of the road ahead. In a step b), at least one signature on the road surface is extracted from the camera images, that is, determined and tracked. From the changed position of the at least one signature in one or more subsequent camera image (s), it is determined in a step c) in which direction the camera is rotated with respect to the roll angle. The amount of roll angle is estimated in a step d). For this purpose, either in a step d1) taking into account the vehicle speed and an imaging model of the camera, the roll angle is estimated directly or in step d2) the roll angle is iteratively increased or decreased by a predetermined correction angle until the roll angle sufficiently compensates for the rotation of the camera. This results in the estimated roll angle as the total correction value.

Furthermore, it is known to determine the roll angle by a completely video-based self-motion estimation and a modeling of a static environment by using a so-called "structure from motion" method. In addition, the roll angle can be determined by a position sensor.

It is the object of the present invention to provide a method and a system of the type mentioned, which allow a determination of the roll angle with little computational effort and with low equipment costs.

The object is solved by the subject matters of the independent claims. Advantageous embodiments are subject of the dependent claims, the following description and the figures.

It is an essence of the invention to extract at least one optical flow from an image sequence or from a video sequence of an environment of a bicycle and, based on the optical flow, to create a roll angle of a camera mounted on the bicycle which records the image sequence and thus also of the two-wheeler to determine.

According to a first aspect of the invention, there is provided a method of detecting a roll angle of a two-wheel during cornering. The method comprises the steps of recording an image sequence of an environment of the bicycle by means of a camera mounted on the bicycle, extracting one or at least one optical flow from the image sequence, determining the roll angle of the camera based on the extracted optical flow, and determining the roll angle of the bicycle using the determined roll angle of the camera.

The method according to the invention makes it possible to calculate the roll angle in a closed method from the at least one optical flow of at least one arbitrary point in space, without an iteration being necessary. At least the optical flow is needed for an image that contains several flow vectors. Thus, the method can be performed with very little computational effort. Furthermore, in particular no additional hardware, for example in the form of yaw rate sensors or position sensors, is necessary for carrying out the method.

According to one embodiment, it is provided that the optical flow is decomposed into an optical translation flow and a rotational optical flow, wherein the roll angle of the camera is determined on the basis of a direction of the optical rotation flow. For the determination of the roll angle of the camera, the following explained assumptions or simplifications are made in order to allow a particularly simple and efficient calculation of the roll angle of the two-wheeler.

Thus, for the sake of simplification, it is assumed that the optical translation flow during a movement of the camera in the direction of an optical axis of the camera is given by vectors which point in the radial direction away from an optical center of the camera. A length of the vectors depends on corresponding 3d world points and is thus initially unknown. Furthermore, it is assumed that an optical flow during a rotation of the camera around its vertical axis yields approximately horizontally extending flow vectors of equal length. Further, it is assumed that a tilted angle flux results when the camera is tilted about the optical axis and then rotated around the vertical axis of the camera as described above.

In addition, it is assumed that an overall movement of the tilted camera is composed of a translation of the camera in the direction of the optical axis and a rotation of the camera about its vertical axis or about an axis transverse to a roadway on which the bicycle is traveling, wherein a translation the camera is neglected across the optical axis. Thus, the optical flux can be composed of the tilt-tilted rotational flux and the translational flux in the direction away from the optical center, and it is further assumed that the angle of inclination of the two-wheel during cornering remains the same, and thus the vector of optical rotation flux for all positions in the picture is approximately identical in direction and amount. The angle of inclination of the optical flow of rotation can correspond to the inclination angle of the camera and the angle of inclination of the two-wheeled vehicle.

The direction of the optical flow of rotation can be calculated by solving a system of equations. By assuming that the at least one optical translation flow extends radially outward from the optical center of the camera, the direction of the translation flow is known, but not its magnitude. Furthermore, the further assumption that the angle of inclination of the two-wheeler during cornering remains the same and thus the vector of the optical rotation flux for all positions in the image is approximately identical in direction and magnitude, not yet the specific direction and the concrete amount of optical rotation flow known ,

It is therefore intended to calculate the magnitude of the at least one extracted translational flux as well as the direction and the magnitude of the optical rotational flow by solving a mathematical equation system. The mathematical equation system can be solved in various known ways, e.g. also iterative.

According to a second aspect of the invention, there is provided a system for detecting a roll angle of a two-wheel during cornering. The system can be communicatively connected to a driver assistance system of the two-wheeled vehicle, so that in particular roll angles determined by the system can be transmitted to the driver assistance system. Furthermore, the system can also be integrated into the driver assistance system. The system according to the second aspect of the invention comprises a camera and a control unit, wherein the camera is adapted to capture an image sequence of an ambience of the bicycle when the camera is mounted on the bicycle and wherein the control unit is adapted to perform an optical flow extract the image sequence, determine the roll angle of the camera based on the extracted optical flow, and determine the roll angle of the two-wheel on the basis of the determined roll angle of the camera.

According to one embodiment, it is provided that the control unit is set up to divide the optical flow into an optical translation flow and a rotational optical flow and to determine the roll angle of the camera based on a direction of the optical rotation flow.

Furthermore, the control unit can be set up to calculate the direction of the optical rotation flow by solving a system of equations.

According to a third aspect of the invention, a bicycle is provided, in particular a motorcycle, wherein the bicycle comprises a system according to the second aspect of the invention.

According to a fourth aspect of the invention, the use of optical flow extracted from an image sequence for determining a roll angle of a two-wheel during cornering is proposed.

• 1 eine Vorderansicht eines Zweirads mit einer Kamera,
• 2 eine Draufsicht auf die Kamera des Zweirads während einer Kurvenfahrt des Zweirads,
• 3 eine Darstellung eines optischen Rotationsflusses bei einer Rotation der Kamera nach 2,
• 4 eine Darstellung eines optischen Rotationsflusses bei einer Rotation der um ihre optische Achse gekippten Kamera nach 2,
• 5 eine Darstellung eines optischen Translationsflusses bei einer Translationsbewegung der Kamera nach 2,
• 6 eine Darstellung von optischen Gesamtflüssen, welche aus einer Videosequenz extrahiert wurden,
• 7 ein Flussdiagramm eines Ausführungsbeispiels eines Verfahrens zur Ermittlung eines Rollwinkels eines Zweirads während einer Kurvenfahrt.

Embodiments of the invention are explained in more detail below with reference to the schematic drawing. This shows

• 1 a front view of a two-wheeler with a camera,
• 2 a top view of the camera of the two-wheeler during a cornering of the bicycle,
• 3 a representation of an optical flow of rotation during a rotation of the camera after 2 .
• 4 a representation of an optical flow of rotation during a rotation of the camera tilted about its optical axis 2 .
• 5 a representation of an optical translation flow during a translational movement of the camera 2 .
• 6 a representation of total optical fluxes extracted from a video sequence,
• 7 a flowchart of an embodiment of a method for determining a roll angle of a bicycle during cornering.

1 shows a two-wheeler 1 in the embodiment shown in the form of a motorcycle. The bicycle 2 drives in an inclined position on a level road 2 , Slope in this context is to be understood that an inclination angle φ between the road 2 and a vertical axis H of the two-wheeler is less than 90 °. The line N represents a road normal. An angle between the road normal N and the vertical axis H is the roll angle λ of the two-wheeler 1 ,

In the embodiment shown, the vertical axis H of the bicycle falls 1 with a vertical axis H of a camera mounted on the bicycle 3 together, but this is not mandatory. Therefore, in the illustrated embodiment, the inclination angle φ and the roll angle λ of the two-wheeler 1 an inclination angle φ and a roll angle λ of the camera 3 , Unless the Hochachsen H of two-wheeler 1 and camera 3 do not coincide and do not run parallel to one another, the inclination angle φ and the roll angle λ can be calculated correspondingly via an angular difference of the vertical axes H.

The camera 3 can be a control unit 6 and can be attached to the two-wheeler 1 be mounted such that the camera 3 in the event of a steering wheel impact 4 of the two-wheeler 1 around the vertical axis H of the two-wheeler 1 is pivoted about the respective steering angle, but this is not mandatory. Alternatively, the camera, for example, on a fixed frame of the bicycle 1 be attached. Unless the camera 3 the steering angle of the handlebar 3 can follow that is made possible by means of the camera 3 Image recordings in a direction intended by the respective steering angle of the two-wheeler 1 be recorded.

2 shows the bicycle 1 or the camera 3 of the two-wheeler 1 to 1 at a first time (in 2 shown below) and to a second second point (in 2 shown above) during cornering. By means of the camera 3 becomes while cornering the two-wheeler 1 a video sequence of an environment of the two-wheeler 1 added. An area of the recorded environment is determined in particular by an optical axis L of the camera and its orientation, wherein the alignment in particular by the respective steering angle and the inclination angle φ of the bicycle 1 is dependent. In particular, during cornering, a first image and a second image of the environment can be recorded.

At the first time and at the second time is the two-wheeler 1 in cornering. At the first time, the vertical axis H of the camera runs 3 parallel to the road normal N. To take the turn, the steering wheel 4 smashed, leaving the camera 3 during cornering in response to the steering angle around the vertical axis H of the camera 3 or to the road normal N is twisted. It is assumed that an optical flow in such a rotation results in approximately horizontally extending flow vectors of the same length, as by 3 is shown.

Further, it is assumed that the camera 3 in addition to the rotation about the vertical axis H of the camera 3 is tilted about the optical axis L depending on the inclination angle φ. It is assumed that there is a tilted by the inclination angle φ optical flow, as by 4 is shown. The inclination angle φ of the tilted optical flux corresponds to the inclination angle φ of the camera 3 or the two-wheeler 1 while cornering.

For the determination of the inclination angle φ and the roll angle λ is simplified assuming that during cornering of the two-wheeler 1 an overall movement of the tilted about the optical axis L camera 3 from the individual below, through 2 composed movements.

On the one hand experienced by the tilted about the optical axis L camera 3 a translation Δx in the direction of the optical axis L. It is assumed for simplicity that a corresponding optical translation flow during a movement of the camera 3 in the direction of the optical axis L is given by vectors which extend in the radial direction from an optical center 5 the camera 3 show it out like this 5 is shown.

Furthermore, the camera learns 3 a translation Δy transverse to the optical axis L. The translation Δy is simplified for the calculation of the inclination angle φ and the roll angle λ not taken into account.

bzw. $\overrightarrow{p_2}$

ergibt sich somit aus einer vektoriellen Addition eines Translationsflusses (gezeigt durch 5) sowie eines Rotationsflusses (gezeigt durch 4) der um die optische Achse L gekippten Kamera 3.In addition, the camera tilted about the optical axis L experiences 3 a rotation Δα about the vertical axis H of the camera 3 or about the road normal N. A corresponding optical rotation flow is through 4 shown. A simplifying assumed total flow $\stackrel{}{p_{}}$$\overrightarrow{{}_1}$

respectively. $\overrightarrow{p_2}$

thus results from a vectorial addition of a translation flow (shown by 5 ) and a rotational flow (shown by 4 ) of the tilted about the optical axis L camera 3 ,

und $t_2\cdot \overrightarrow{e_{t2}}$

sowie Rotationsflüssen $\overrightarrow{r_1}$

und $\overrightarrow{r_2},$

woraus ein jeweiliger Gesamtfluss $$\overrightarrow{p_1}=t_1\cdot \overrightarrow{e_{t1}}+\overrightarrow{r_1}\text{ und }\overrightarrow{p_2}=t_2\cdot \overrightarrow{e_{t2}}+\overrightarrow{r_2}$$

$$\text{mit }\overrightarrow{r_1}=\overrightarrow{r_2}\approx \overrightarrow{r}=const.$$

resultiert. Die Gesamtflüsse $\overrightarrow{p_1}$

und $\overrightarrow{p_2}$

können aus dem ersten aufgenommenen Bild und aus dem zweiten aufgenommenen Bild extrahiert werden. 6 shows two overlays or vectorial additions of translational flows as an example for two pixels $t_{}$${}_1\cdot \overrightarrow{{\mathrm{<figure-callout\; id="1"\; label="e\; t"\; filenames="DE102016220559A1\_0065.png"\; state="\{\{state\}\}">e\; </figure-callout>}}_t}$

and $t_2\cdot \overrightarrow{e_{t2}}$

as well as rotational flows $\stackrel{}{r_{}}$$\overrightarrow{{}_1}$

and $\overrightarrow{r_2}.$

from which a respective total flow $$\stackrel{}{p_{}}$$$$\overrightarrow{{}_1}={\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="DE102016220559A1\_0065.png"\; state="\{\{state\}\}">t</figure-callout>}}_1\cdot \overrightarrow{{\mathrm{<figure-callout\; id="1"\; label="e\; t"\; filenames="DE102016220559A1\_0065.png"\; state="\{\{state\}\}">e\; </figure-callout>}}_t}+\overrightarrow{{\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="DE102016220559A1\_0065.png"\; state="\{\{state\}\}">r</figure-callout>}}_1}\text{and}\overrightarrow{p_2}=t_2\cdot \overrightarrow{e_{t2}}+\overrightarrow{r_2}$$

$$\text{With}\overrightarrow{r_1}=\overrightarrow{r_2}\approx \overrightarrow{r}=cOnst,$$

results. The total flows $\overrightarrow{{\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="DE102016220559A1\_0065.png"\; state="\{\{state\}\}">p</figure-callout>}}_1}$

and $\overrightarrow{p_2}$

can be extracted from the first recorded image and from the second captured image.

7 shows the sequence of an embodiment of a method for determining the roll angle A of the two-wheeler 1 to 1 while cornering 2 ,

In a first step 100 becomes while cornering the two-wheeler 1 by means of the camera 3 recorded a video sequence which comprises a first image recorded at a first time t ₁ and a second image recorded at a later time t ₂ . The video sequence is sent to the control unit 6 of the two-wheeler 1 transmitted, wherein the control unit 6 - like through 2 shown - for example, on or in the camera 3 which is not mandatory.

und $\overrightarrow{p_2}.$

In a second step 200 is extracted from the first image and from the second image at least one total flow in which 6 shown two total optical fluxes $\stackrel{}{p_{}}$$\overrightarrow{{}_1}$

and $\overrightarrow{p_2},$

This extraction can be done by means of the control unit set up for this purpose 6 respectively.

und $t_2\cdot \overrightarrow{e_{t2}}$

von einem optischen Zentrum 5 der Kamera 3 aus radial nach außen verlaufen, wodurch zwar die Richtungen $\overrightarrow{e_{t1}}$

und $\overrightarrow{e_{t2}}$

der Translationsflüsse $t_1\cdot \overrightarrow{e_{t1}}$

und $t_2\cdot \overrightarrow{e_{t2}}$

bekannt sind, nicht jedoch deren Längen bzw. Beträge t₁ und t₂. Weiterhin wird angenommen, dass der Neigungswinkel φ des Zweirads 1 während der Kurvenfahrt gleich bleibt und somit der Vektor des optischen Rotationsflusses $\overrightarrow{r}$

für alle Positionen im Bild näherungsweise identisch in Richtung und Betrag ist. Allerdings sind die konkrete Richtung und der konkrete Betrag des optischen Rotationsflusses $\overrightarrow{r}$

zunächst unbekannt.In a third step 300 becomes the inclination angle φ of the camera 3 calculated, which also by means of the dedicated control unit 6 can be done. It is assumed that the optical translation flows ${\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="DE102016220559A1\_0065.png"\; state="\{\{state\}\}">t</figure-callout>}}_1\cdot \overrightarrow{{\mathrm{<figure-callout\; id="1"\; label="e\; t"\; filenames="DE102016220559A1\_0065.png"\; state="\{\{state\}\}">e\; </figure-callout>}}_t}$

and $t_2\cdot \overrightarrow{e_{t2}}$

from an optical center 5 the camera 3 from radially outward, whereby the directions $\stackrel{}{e_t}$$\overrightarrow{1_{}}$

and $\overrightarrow{e_{t2}}$

the translation flows ${\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="DE102016220559A1\_0065.png"\; state="\{\{state\}\}">t</figure-callout>}}_1\cdot \overrightarrow{{\mathrm{<figure-callout\; id="1"\; label="e\; t"\; filenames="DE102016220559A1\_0065.png"\; state="\{\{state\}\}">e\; </figure-callout>}}_t}$

and $t_2\cdot \overrightarrow{e_{t2}}$

are known, but not their lengths or amounts t ₁ and t ₂ . Furthermore, it is assumed that the inclination angle φ of the two-wheeler 1 while cornering remains the same and thus the vector of the optical flow of rotation $\overrightarrow{r}$

for all positions in the picture is approximately identical in direction and amount. However, the concrete direction and the concrete amount of the optical rotation flow are $\overrightarrow{r}$

initially unknown.

und $t_2\cdot \overrightarrow{e_{t2}}$

berechnet. Ferner werden auch die Beträge t₃ bis t_i weiterer, nicht durch 6 gezeigter extrahierter optischer Gesamtflüsse $\overrightarrow{p_3}$

bis $\overrightarrow{p_l}$

berechnet. Dies kann durch die dazu eingerichtete Steuerungseinheit 6 erfolgen. Dabei werden auch die Richtung und der Betrag des optischen Rotationsflusses $\overrightarrow{r}$

berechnet.Therefore, by solving a mathematical equation system described below, the amounts t ₁ and t _{2 of} the translation flows become ${\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="DE102016220559A1\_0065.png"\; state="\{\{state\}\}">t</figure-callout>}}_1\cdot \overrightarrow{{\mathrm{<figure-callout\; id="1"\; label="e\; t"\; filenames="DE102016220559A1\_0065.png"\; state="\{\{state\}\}">e\; </figure-callout>}}_t}$

and $t_2\cdot \overrightarrow{e_{t2}}$

calculated. Furthermore, the Amounts t ₃ to t _i further, not by 6 shown extracted total optical fluxes $\stackrel{}{p_{}}$$\overrightarrow{{}_3}$

to $\overrightarrow{p_l}$

calculated. This can be done by the dedicated control unit 6 respectively. At the same time the direction and the amount of the optical rotation flow become $\overrightarrow{r}$

calculated.

und $\overrightarrow{p_2}$

können jeweils in eine x-Komponente und in eine y-Komponente zerlegt werden: $$p_{1x}=t_1\cdot e_{t1x}+r_x,$$

$$p_{1y}=t_1\cdot e_{t1y}+r_y,$$

$$p_{2x}=t_2\cdot e_{t2x}+r_x\text{ und}$$

$$p_{2y}=t_2\cdot e_{t2y}+r_y.$$

The total optical fluxes $\stackrel{}{p_{}}$$\overrightarrow{{}_1}$

and $\overrightarrow{p_2}$

can each be decomposed into an x-component and a y-component: $$p_{1x}={\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="DE102016220559A1\_0065.png"\; state="\{\{state\}\}">t</figure-callout>}}_1\cdot e_{t1x}+r_x.$$

$$p_{1y}={\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="DE102016220559A1\_0065.png"\; state="\{\{state\}\}">t</figure-callout>}}_1\cdot e_{t1y}+r_y.$$

$$p_{2x}=t_2\cdot e_{t2x}+r_x\text{and}$$

$$p_{2y}=t_2\cdot e_{t2y}+r_y,$$

bis $\overrightarrow{p_l}$

erfolgen: $$\begin{array}{l}{P}_{3x}=t_3\cdot e_{t3x}+r_x,\hfill \\ P_{3y}=t_3\cdot e_{t3y}+r_y,\hfill \\ ⋮\hfill \\ P_{ix}=t_i\cdot e_{tix}+r_x\text{ und}\hfill \\ P_{iy}=t_i\cdot e_{tiy}+r_y.\hfill \end{array}$$

The same decomposition can also for the further, not through 6 shown extracted total optical fluxes $\stackrel{}{p_{}}$$\overrightarrow{{}_3}$

to $\overrightarrow{p_l}$

respectively: $$\begin{array}{l}{P}_{3x}=t_3\cdot e_{t3x}+r_x.\hfill \\ P_{3y}=t_3\cdot e_{t3y}+r_y.\hfill \\ ⋮\hfill \\ P_{ix}=t_i\cdot e_{tix}+r_x\text{and}\hfill \\ P_{iy}=t_i\cdot e_{tiy}+r_y,\hfill \end{array}$$

This system of equations can be represented in matrix notation as follows: $$\left[\begin{array}{l}{p}_{1x}\hfill \\ p_{1y}\hfill \\ p_{2x}\hfill \\ p_{2x}\hfill \\ ⋮\hfill \end{array}\right]=\left[\begin{array}{lllll}{e}_{t1x}\hfill & 0\hfill & \cdots \hfill & 1\hfill & 0\hfill \\ e_{t1y}\hfill & 0\hfill & \cdots \hfill & 0\hfill & 1\hfill \\ 0\hfill & e_{t2x}\hfill & \cdots \hfill & 1\hfill & 0\hfill \\ 0\hfill & e_{t2y}\hfill & \cdots \hfill & 0\hfill & 1\hfill \end{array}\right]\cdot \left[\begin{array}{l}{t}_1\hfill \\ t_2\hfill \\ ⋮\hfill \\ r_x\hfill \\ {\text{r}}_y\hfill \end{array}\right],$$

The individual matrices can be abbreviated as follows: $$\left[\begin{array}{l}{p}_{1x}\hfill \\ p_{1y}\hfill \\ p_{2x}\hfill \\ p_{2x}\hfill \\ ⋮\hfill \end{array}\right]=b.\left[\begin{array}{lllll}{e}_{t1x}\hfill & 0\hfill & \cdots \hfill & 1\hfill & 0\hfill \\ e_{t1y}\hfill & 0\hfill & \cdots \hfill & 0\hfill & 1\hfill \\ 0\hfill & e_{t2x}\hfill & \cdots \hfill & 1\hfill & 0\hfill \\ 0\hfill & e_{t2y}\hfill & \cdots \hfill & 0\hfill & 1\hfill \end{array}\right]=A\text{and}\left[\begin{array}{l}{t}_1\hfill \\ t_2\hfill \\ ⋮\hfill \\ r_x\hfill \\ {\text{r}}_y\hfill \end{array}\right]=x,$$

$$A^T\cdot b=A^T\cdot A\cdot x$$

$$x={\left(A^T\cdot A\right)}^{-1}{A}^T\cdot b.$$

The equation system b = A ▪ x can be solved for the unknown x in the following way: $$b=A\cdot x$$

$$A^T\cdot b=A^T\cdot A\cdot x$$

$$x={\left(A^T\cdot A\right)}^{-1}{A}^T\cdot b,$$

des optischen Rotationsflusses wird der Neigungswinkel φ der Kamera 3 berechnet. Dabei wird angenommen, dass der Neigungswinkel φ des gekippten optischen Rotationsflusses dem Neigungswinkel φ der Kamera 3 bzw. dem Zweirad 1 während der Kurvenfahrt entspricht. In einem vierten Schritt 400 wird aus dem Neigungswinkel der Kamera 3 der Neigungswinkel φ des Zweirads 1 berechnet, wobei die beiden Winkel in dem durch 1 dargestellten Ausführungsbeispiel eines Zweirads 1 mit der Kamera 3 identisch sind. Falls die beiden Neigungswinkel nicht identisch sind, können bei Kenntnis einer entsprechenden Winkeldifferenz die Neigungswinkel leicht ineinander umgerechnet werden, z.B. mittels der dazu eingerichteten Steuerungseinheit 6.From the calculated or graphically determined direction of the vector $\overrightarrow{r}$

of the optical rotation flux becomes the inclination angle φ of the camera 3 calculated. It is assumed that the inclination angle φ of the tilted optical rotation flux corresponds to the inclination angle φ of the camera 3 or the two-wheeler 1 while cornering. In a fourth step 400 is from the tilt angle of the camera 3 the angle of inclination φ of the bicycle 1 calculated, with the two angles in the by 1 illustrated embodiment of a bicycle 1 with the camera 3 are identical. If the two angles of inclination are not identical, the inclination angles can be easily converted into one another, for example by means of the control unit set up for this purpose, if a corresponding angular difference is known 6 ,

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• EP 1989086 B1 [0002]
• WO 2011/063785 A1 [0003]

Claims (8)

Method for determining a roll angle (λ) of a two-wheeler (1) during cornering, the method comprising the steps of: (100) taking a sequence of images of an environment of the two-wheeled vehicle (1) by means of a camera (3) mounted on the bicycle (1), (200) Extract an optical flow $(\overrightarrow{p_1}$

to $\overrightarrow{p_l})$

from the image sequence, (300) determining the roll angle (λ) of the camera (3) on the basis of the extracted optical flux $(\overrightarrow{p_1}$

to $\overrightarrow{p_l})$

and (400) determining the roll angle (λ) of the two-wheeler (1) on the basis of the determined roll angle (λ) of the camera (3). Method according to Claim 1 where the optical flow $(\overrightarrow{p_1}$

to $\overrightarrow{p_l})$

in an optical translation flow $(t_1\cdot \overrightarrow{e_{t1}}$

to $t_i\cdot \overrightarrow{e_{tl}})$

and in a rotational optical flow $\left(\overrightarrow{r}\right)$

is decomposed, using a direction of optical rotation flow $\left(\overrightarrow{r}\right)$

the roll angle (λ) of the camera (3) is determined. Method according to Claim 2 where the direction of the optical flow of rotation $\left(\overrightarrow{r}\right)$

is calculated by solving a system of equations. System for determining a roll angle (λ) of a two-wheeled vehicle (1) during cornering, the system comprising a camera (3) and a control unit (6), the camera (3) being adapted to record an image sequence of an environment of the two-wheeled vehicle (1) when the camera (3) is mounted on the bicycle (1), and wherein the control unit (6) is adapted to provide an optical flow $(\overrightarrow{p_1}$

to $\overrightarrow{p_l})$

to extract from the image sequence, the roll angle (λ) of the camera (3) based on the extracted optical flux $(\overrightarrow{p_1}$

to $\overrightarrow{p_l})$

to determine, and to determine the roll angle (λ) of the two-wheeler (1) on the basis of the determined roll angle (λ) of the camera (3). System after Claim 4 , wherein the control unit (6) is adapted to the optical flow $(\overrightarrow{p_1}$

to $\overrightarrow{p_l})$

in an optical translation flow $(t_1\cdot \overrightarrow{e_{t1}}$

to $t_i\cdot \overrightarrow{e_{tl}})$

and in a rotational optical flow $\left(\overrightarrow{r}\right)$

to disassemble and by a direction of the optical rotation flow $\left(\overrightarrow{r}\right)$

to determine the roll angle (λ) of the camera (3). System after Claim 5 wherein the control unit (6) is adapted to the direction of the optical flow of rotation $\left(\overrightarrow{r}\right)$

by solving a system of equations. A two-wheeled vehicle (1), in particular a motorcycle, comprising a system according to one of Claims 4 to 6 , Use of optical flow extracted from a sequence of images $(\overrightarrow{p_1}$

to $\overrightarrow{p_l})$

for determining a roll angle (λ) of a two-wheeler (1) during cornering.

Images