Detection method of motor vehicle exit lane

Abstract

The present invention relates to a method for detecting an exit lane for a motor vehicle, the method comprising: detecting ground marking lines on the right side (LR) and the left side (LL) that laterally delimit a main lane (L) in an image acquired by a vehicle-mounted camera, determining a first heading angle and a second heading angle relative to the detected marking lines, providing a first standard exit lane detection state based on a comparison of the first heading angle and the second heading angle with a predetermined threshold, and providing a second coaxial exit lane detection state (EXIT_R) based on a comparison between the derivatives of the changes of the first heading angle and the second heading angle with respect to time and verification that the value of the first heading angle or the value of the second heading angle remains substantially close to a zero value.

Description

[0001] The present invention relates to the field of controlling the trajectory of a motor vehicle by detecting road markings on the ground using a camera, in particular a camera on board the motor vehicle. More particularly, the present invention relates to a method for detecting a motor vehicle approaching an exit lane.

[0002] The invention is advantageously applied to the implementation of systems for managing the level of autonomy of automated motor vehicles, for which it is crucial to be able to reliably control the positioning of the vehicle in its environment. In particular, the problem of controlling the trajectory of a vehicle today is keyed to allowing a first level of autonomy for the vehicle in an automated driving phase, i.e. during which the vehicle is driven in an automated manner without the need for the driver to take action on the steering wheel or pedals.

[0003] Algorithms implemented in a video processing unit that obtains a video stream from a camera on board a vehicle are known, which can extract road marking lines on the ground from the video stream, detect the position of the motor vehicle relative to the road markings on the ground, and correct the trajectory of the vehicle relative to the detected road markings. "Markings" means lines on the ground that are different in color from the road and that delimit one side of the lane. The marking lines on the ground can be continuous or discontinuous.

[0004] However, most of the methods used by the detection algorithms have weaknesses in certain common real-world situations, such as when approaching a so-called exit that is coaxial with respect to the main lane of vehicle travel, i.e., the deceleration lane of the exit is substantially in the continuation of the main lane of vehicle travel. In the present application, "the deceleration lane is substantially in the continuation of the main lane" means that the exit lane is substantially along a tangent to the main lane.

[0005] Nowadays, the detection of the exit lane at the intersection relative to the main lane of the vehicle can be established based on the value of the heading angle of the vehicle relative to the observed line, that is, the angle between the left and right ground markings of the lane and the vehicle reference system. Figure 1 As shown, ψ R is the heading angle of the vehicle 10 relative to the right marking line L1, and ψ L is the heading angle of the vehicle relative to the left marking line L2.

[0006] For most exits, a significant increase in heading angle can be observed on the exit lane side. For example, when the exit starts on the right, the heading angle on the right side increases sharply. R The exit is detected by comparing with a predetermined threshold.

[0007] Conversely, in cases where the exit lane is substantially in continuation of the main lane, the configuration of the road markings may "fool" the detection algorithm, since in these cases the heading angle on one side of the exit does not increase sharply. These cases typically involve deceleration lanes leading to highway exits or deceleration lanes leading to highway service areas.

[0008] Figure 2 An example illustrating this problem is shown. The main lane in which the vehicle is traveling is the right lane D of the road indicated by 1 and is bounded by ground markings 2 and 3 which delimit the main lane of the vehicle laterally on the right and left sides. The lane turns substantially to the left, and there is a deceleration lane 4 in the immediate vicinity in front of the vehicle, which is substantially in the continuation of the main lane D of the vehicle relative to the current position of the vehicle in its lane. The exit line 5 delimiting the deceleration lane 4 on the right side on the ground is substantially straight, while the ground markings 3 delimiting the main lane D on the right side are substantially curved to the left. In other words, here, the detection algorithm processing the data provided by the on-board camera should generally extract information about the road markings of the main lane, i.e., the ground markings 2 and 3, and determine the corrective action to be taken so that the vehicle trajectory is effectively based on the detection of the ground markings 2 and 3, so that the vehicle remains in the main lane. However, due to the configuration of the ground markings of the main lane D and the deceleration lane 4 (which is substantially in the continuation of the main lane), the detection algorithm may erroneously detect the exit line 5 of the deceleration lane 4 as the ground marking of the main lane. Since this line is mistakenly detected as a ground marking line of the main lane, the corrective action applied to the vehicle trajectory will cause the vehicle to leave the main lane and enter the deceleration lane, which is not the desired action.

[0009] Document US 2018156626 discloses a system that controls the trajectory of a vehicle along a "normal" route by using an on-board camera to detect and follow the ground marking lines of the lane in which the vehicle is traveling, and when approaching a highway entrance or exit lane, determining whether the ground marking lines should still be followed based on the "normal" route that the vehicle will follow. More precisely, the detection of the ground marking lines is based on information about the color of the ground marking lines. Therefore, based on the color information of the predetermined ground marking line corresponding to the "normal" driving route, if the color of the detected ground marking line is the same as the color information of the predetermined ground marking line, then continue to follow the detected line, otherwise interrupt.

[0010] Therefore, the system should a priori be robust to line detection errors associated with approaching coaxial exit lanes, since the detection of ground-marking lines is not based on angle measurements. However, the detection here relies on the analysis of the color of the detected ground-marking lines, which is not without disadvantages. In particular, the compliance and recognition of road marking lines may depend on the acquisition conditions of the on-board image, which is easily affected by cast shadows, glare, etc., and is therefore difficult to use in terms of fine colors.

[0011] The object of the present invention is therefore to provide a method for detecting the approach of an exit lane which is in particular able to detect the approach of an exit lane coaxial with the vehicle's main lane in a simple and robust manner so that the position of the vehicle can be corrected if necessary to keep it in its lane.

[0012] To this end, the present invention relates to a method for detecting an exit lane of a motor vehicle based on a series of images in front of the vehicle, these images being acquired by a camera on board the vehicle while the vehicle is moving along a main carriageway, the main carriageway being laterally delimited by ground markings located along the right and left edges of said carriageway, the method comprising:

[0013] a step of detecting a right ground marking line and a left ground marking line in the image,

[0014] The step of determining a first heading angle of the vehicle relative to a detected right ground marking line and a second heading angle of the vehicle relative to a detected left ground marking line,

[0015] A step of providing a first exit lane detection state according to a comparison between the value of the first heading angle or the value of the second heading angle and a predetermined threshold, wherein the first exit lane detection state represents that a right standard exit lane or a left standard exit lane is detected, respectively;

[0016] The method is characterized in that the method further includes a step of providing a second exit lane detection state based on firstly comparing the derivatives of the changes of the first heading angle and the second heading angle with respect to time and secondly verifying that the value of the first heading angle or the value of the second heading angle respectively remains substantially close to a zero value, the second exit lane detection state representing the detection of a right exit lane or a left exit lane respectively coaxial with the main driving lane.

[0017] Thus, by performing a validation test on the variation of these two heading angles of the vehicle relative to the ground markings detected on the right and on the left, the second detection state advantageously enables the detection of so-called coaxial exit lanes relative to the main carriageway, which are not taken into account by the first detection state, which is limited to so-called standard exit lanes, for which a simple increase of the heading angle of the vehicle on the side of the exit relative to a given threshold can be observed. Thus, in the present application, the expression "standard exit" refers to an exit that is not coaxial with the main carriageway, i.e. without a deceleration lane in the continuation of the main lane in which the vehicle travels.

[0018] Thus, in the case of ambiguous ground markings, typically for exit lanes coaxial with the main traffic lane, the method of the invention is able to detect them in a robust manner with the aim of correcting the position of a vehicle that mistakenly enters such an exit lane.

[0019] The exit lane detection according to the invention is applicable to both right and left exit lanes, which provides a great flexibility of solution for any type of vehicle, both in countries with legislation for driving on the left and in countries with legislation for driving on the left.

[0020] Advantageously, the method comprises arbitrating between the first exit lane detection state and the second exit lane detection state to grant processing priority to the first detection state when the second detection state has been provided.

[0021] Advantageously, when the derivative of the change of the second heading angle or the first heading angle with respect to time is respectively greater than the derivative of the change of the first heading angle or the second heading angle with respect to time plus a predefined minimum margin, and the value of the first heading angle or the value of the second heading angle is respectively zero to a predetermined tolerance threshold, the second exit lane detection status is provided for the right coaxial exit lane or the left coaxial exit lane, respectively.

[0022] Advantageously, when the value of the first heading angle or the value of the second heading angle is respectively higher than the predetermined threshold, the first exit lane detection state is provided for the right standard exit lane or the left standard exit lane respectively.

[0023] Advantageously, the method comprises the step of correcting the position of a ground marking line detected on a side of the exit lane detected according to the first exit lane detection state or the second exit lane detection state.

[0024] Advantageously, correction of the ground marking lines detected on one side of the detected exit lane comprises reconstructing corrected ground marking lines on one side of the detected exit lane based on the positions of the ground marking lines detected on the opposite side relative to the detected exit lane and the measured values ​​of the lane width obtained using the camera.

[0025] Advantageously, said measured value of the lane width comprises a measured value fixed to a measured value taken at a time before said exit lane was detected.

[0026] Advantageously, the method comprises the step of correcting the lateral position of the vehicle in its lane based on the position of the detected ground markings on the opposite side relative to the detected exit lane and the position of the corrected ground markings on the side of the detected exit lane.

[0027] The present invention also relates to a device for detecting an exit lane for a motor vehicle traveling on a main lane laterally delimited by ground markings, the device being characterized in that the device comprises at least one processing unit suitable for cooperating with a front camera equipped on the vehicle and for implementing the steps of the above method.

[0028] The invention also relates to a motor vehicle, characterised in that it comprises at least one on-board front camera and a device as described above.

[0029] Other characteristics and advantages of the invention will become more apparent from a reading of the following description given by way of illustrative and non-limiting example and with reference to the following single drawing:

[0030] [ Figure 1 ] schematically illustrates the concept of the vehicle's heading angle relative to ground markings on the right and left sides;

[0031] [ Figure 2 ] schematically illustrates the problem of ambiguous ground markings of the so-called coaxial exit lane, misleading the camera-based line detection system;

[0032] [ Figure 3 ] schematically illustrates the principle of detecting an exit lane in the context of detecting a right coaxial exit lane;

[0033] [ Figure 4 ] schematically illustrates the principle of detecting an exit lane in the context of detecting a left coaxial exit lane;

[0034] [ Figure 5 ] schematically illustrates the principle of detecting an exit lane against the background of detecting a left standard exit lane;

[0035] [ Figure 6] schematically illustrates the principle of detecting the exit lane against the background of detecting the right standard exit lane.

[0036] Figure 3 The diagram schematically shows that the vehicle 10 is moving along a road marking line (ie, a right ground marking line L located along the right edge of the main lane L). R and the left ground marking line L located along the left edge of the main carriageway L L ) An exemplary situation of driving on a laterally defined main lane L.

[0037] The vehicle 10 is equipped with an assistance system generally referred to as "lane centering assist" or by the acronym LCA, which consists in detecting the position of the motor vehicle relative to the road markings and correcting the trajectory of the vehicle so as to automatically keep the vehicle in the center of its lane, thereby relieving the driver of the steering task. Thus, when a trajectory correction is required, the LCA system acts on the steering column of the vehicle.

[0038] An LCA system typically includes a camera facing forward of a motor vehicle, such as mounted on top of a windshield, at its junction with the roof of the vehicle, or behind an interior rearview mirror of the vehicle, so that images of the scene in front of the vehicle can be acquired and road markings in front of the motor vehicle can be detected.

[0039] To this end, the step of acquiring images by means of the camera is followed by a step in which a processing unit associated with the camera takes the image stream provided by the camera and analyses it using image processing techniques, in particular to detect road marking lines located along the lane L occupied by the motor vehicle 10. The processing unit is also suitable for providing estimates of two heading angles of the vehicle relative to the road marking lines detected on either side of the vehicle.

[0040] Figure 3 The bifurcation of the main lane is shown, including a right-hand exit lane EXIT_R, which extends in continuation of the main lane L of the vehicle, while the main lane branches off to the left at the bifurcation. The right-hand exit lane is a so-called exit lane coaxial with the main lane, i.e., an exit lane which is substantially in continuation of the main lane. This coaxial exit lane EXIT_R itself is laterally delimited by road markings, i.e., the right-hand ground marking line EXIT_R is positioned along the right-hand edge of the right-hand coaxial exit lane EXIT_R. R and the left ground marking line EXIT positioned along the left edge of the right coaxial exit lane EXIT_R L .

[0041] Figure 3Such a bifurcation situation as shown in the figure generally describes a configuration of blurred ground markings that can easily "fool" the detection algorithm typically used by the camera, which is based on the comparison of the vehicle heading angle value with a predetermined threshold. Therefore, at the beginning of the detection, when approaching the right coaxial exit lane EXIT_R, the onboard camera may mistakenly identify the right ground marking line EXIT_R, which is located along the right edge of the right coaxial exit lane EXIT_R. R Detected as ground markings of the main lane L, this may result in applying unnecessary corrective actions to the vehicle trajectory, causing it to take the right coaxial exit lane EXIT_R, when in fact it should remain in the left-turning main lane L. In other words, this type of coaxial exit lane is not detected by the conventional detection algorithm associated with the camera, thereby erroneously treating this type of coaxial exit lane as the main lane of the vehicle.

[0042] In this regard, the present invention provides for implementing a second detection algorithm designed to specifically detect this type of coaxial exit lane, such as according to Figure 3 In order to be able to correct the trajectory of the vehicle to keep it in its main lane. Typically, such detection algorithms are designed to detect coaxial exit lanes by firstly comparing the average derivatives of the two heading angles on the right and left sides over a certain period of time and secondly verifying that the vehicle's heading angle relative to the road markings detected on one side of the exit lane is indeed zero.

[0043] Therefore, by applying this general principle to Figure 3 In the example of , the coaxial exit lane detection algorithm can provide a detection state representing the detection of the right coaxial exit lane EXIT_R when the following two cumulative conditions are met: that is, the increase in the left heading angle is greater than the right heading angle and the right heading angle remains close to zero to within a predetermined tolerance threshold. This last condition that the right heading angle must remain close to zero can eliminate valid detection situations of the left exit lane, in which case the right heading angle will not remain close to zero.

[0044] This dual test implemented by the coaxial exit lane detection algorithm to determine whether a vehicle is approaching a right coaxial exit lane may take the form of the following dual expression:

[0045] [Math.1]

[0046]

[0047] ψ R (k)<ε (1)

[0048] in:

[0049] nDelay: indicates the left heading angle ψ Land the right heading angle ψ R The number of time intervals of the derivative period;

[0050] k: time interval;

[0051] Margin: safety margin to avoid untimely detection;

[0052] ε: A predetermined tolerance threshold used to test for zero values ​​of the right heading angle.

[0053] Therefore, the above criterion (1) can distinguish the right coaxial exit lane from the left coaxial exit lane. Specifically, for the right coaxial exit lane, the right heading angle ψ R will be very low or even zero. Therefore, it is verified that its value is below a predetermined tolerance threshold ε.

[0054] Just refer to Figure 3 The described principle for detecting the right coaxial exit lane benefits from a perfectly symmetrical application for detecting the left coaxial exit lane, e.g. Figure 4 as shown in the example.

[0055] therefore, Figure 4 The fork shown in FIG. 1 comprises a left exit lane EXIT_L which extends in the continuation of the main traffic lane L of the vehicle, while the main traffic lane branches off to the right at the fork. The left coaxial exit lane EXIT_L itself is laterally delimited by road marking lines, i.e., a right ground marking line EXIT_L positioned along the right edge of the left coaxial exit lane EXIT_L. R and the left ground marking line EXIT located along the left edge of the left coaxial exit lane EXIT_L L .

[0056] Thus, the coaxial exit lane detection algorithm provides a detection state representing detection of the left coaxial exit lane EXIT_L when the following two cumulative conditions are met: namely, the increase in the right heading angle of the vehicle is greater than the left heading angle and the left heading angle remains close to zero to within a predetermined tolerance threshold.

[0057] In this case, the dual test implemented by the detection algorithm to determine if the left coaxial exit lane is being approached takes the following form:

[0058] [Math.2]

[0059]

[0060] ψ L (k)<ε (2)

[0061] As in the previous case, criterion (2) can distinguish the right coaxial exit lane from the left coaxial exit lane. Specifically, for the left coaxial exit lane, the left heading angle ψ L will be very low or even zero. Therefore, it is verified that its value is below a predetermined tolerance threshold ε.

[0062] Therefore, by applying criteria (1) and (2), the detection of the left coaxial exit lane and the right coaxial exit lane cannot coexist. Specifically, it is impossible for the right heading angle and the left heading angle to be zero at the same time and detect the exit lane at the same time.

[0063] Therefore, the camera is associated with a first detection algorithm which is generally intended to detect an exit lane based on a comparison of the values ​​of the right and left heading angles of the vehicle, respectively, with predetermined thresholds. Specifically, for most "standard" exit lanes, a clear increase in heading angle can be observed relative to the line detected on one side of the exit lane. Therefore, the algorithm is able to provide a first detection state representing a "standard" detection, rather than a coaxial exit lane on the right or left side. However, as Figure 3 and Figure 4 As shown in the example of , the first detection algorithm is unable to detect the coaxial exit lanes on the right and left sides respectively, and therefore, a second algorithm implemented as described above is provided for this purpose.

[0064] In other words, the controller of the LCA system (which is intended to correct the vehicle's trajectory to keep the vehicle in its lane when encountering a fork in the road) processes information from multiple sources (corresponding to the detection states provided by the first detection algorithm and the second detection algorithm) to detect an exit lane.

[0065] Advantageously, provision is made for arbitration between these exit lane detection states so as to grant processing priority to the detection state provided by the first detection algorithm. Such arbitration giving priority to the detection state provided by the first detection algorithm may avoid regression in already existing correction systems based on information provided by the first detection algorithm for detecting standard exit lanes on the right and on the left. In particular, giving priority to the detection state provided by the first detection algorithm may avoid a situation where a false detection ends up occurring, i.e., a coaxial exit lane on the right or left, respectively, is falsely detected instead of a standard exit lane on the left or right, respectively.

[0066] Figure 5 A fork configuration is described to demonstrate the advantages of such arbitration. The fork junction here comprises a left standard exit lane EXIT_STD_L relative to the main traffic lane L of the vehicle, on one side of which a left heading angle can be observed relative to a road marking line EXIT positioned along the left edge of the left standard exit lane EXIT_STD_L. LIn this case, the algorithm for detecting standard exit lanes therefore does provide a detection state representing that the left standard exit lane EXIT_STD_L is detected. At the same time, the coaxial exit lane detection algorithm also provides a detection state representing that the right coaxial exit lane is detected because both of the following conditions are met: namely, the left heading angle relative to the road marking line EXIT_STD_L is L The increase exceeds the right heading angle relative to the road marking line L R , and the right heading angle remains very close to zero. However, this last detection state is not considered because it has a lower priority than the detection state of the left standard exit lane. Therefore, arbitration makes it possible to consider detections according to a priority scheme that first considers detections generated by the first algorithm for detecting the standard exit lane, and then considers detections generated by the second algorithm for detecting the coaxial exit lane only when there is no detection state provided by the first detection algorithm (in other words, when the standard exit lane is not detected).

[0067] The arbitration principle between these two types of detection applies in the same way, so that it is avoided that a left coaxial exit lane is detected when a right standard exit lane should be detected.

[0068] therefore, Figure 6 A bifurcated configuration is described, where a standard exit lane on the right should be detected so that the trajectory can be corrected correctly.

[0069] The junction here comprises a right standard exit lane EXIT_STD_R relative to the main lane L of the vehicle, so that on one side a right heading angle relative to the road marking line EXIT positioned along the right edge of the right standard exit lane EXIT_STD_R can be observed. R In this case, the algorithm for detecting standard exit lanes therefore does provide a detection state representing that the right standard exit lane EXIT_STD_R is detected. At the same time, the coaxial exit lane detection algorithm also provides a detection state representing that the left coaxial exit lane is detected because both of the following conditions are met: namely, the right heading angle relative to the road marking line EXIT_STD_R is R The increase exceeds the left heading angle relative to the road marking line L L The left heading angle remains very close to zero. However, this last detection state is not considered by arbitration because it has a lower priority than the detection state of the right standard exit lane.

[0070] This priority scheme resulting from the implemented arbitration (which provides that detections produced by the first detection algorithm are considered first, and then detections produced by the second detection algorithm are considered only if the first algorithm does not detect anything) can take the following form, for example, in the case of a right exit lane detection, where DETECTION_CAMERA_RIGHT corresponds to the detection state provided by the first standard exit lane detection algorithm, and DETECTION_COAXIAL_RIGHT corresponds to the detection state provided by the second coaxial exit lane detection algorithm:

[0071] Detection_right = Detection_camera_right OR [Detection_coaxial_right AND NOT (Detection_camera_right)]

[0072] In this way, information for detecting coaxial exit lanes is processed as additional information so that coaxial exit lanes that were not returned by the first detection algorithm can be detected.

[0073] Now that the right or left exit lane detection has been made robust in all of these bifurcations, the position of the road markings will be corrected to ensure that the vehicle remains within its main lane. To do this, the ground markings detected on the side where the exit lane is detected will be corrected by reconstructing the corrected ground markings on the side where the exit lane is detected based on the positions of the ground markings detected on the opposite side relative to the detected exit lane and the measurements of the lane width acquired using the camera.

[0074] Therefore, when the image processing unit associated with the camera performs an image processing of the road marking lines L laterally delimiting the main lane L, R and L L When testing, the ground marking line L R and L L The respective lateral position relative to the camera in the vehicle reference frame is also determined.

[0075] Order pos R and pos L are the respective lateral positions of these ground markers relative to the camera in the vehicle reference frame.

[0076] Based on these respective lateral positions, the width of the main lane L is calculated. Let LW(k) be the lane width calculated at time k based on the absolute values ​​of these respective lateral positions:

[0077] [Math.3]

[0078] LW(k)=abs(pos R )+abs(pos L )

[0079] Once an exit lane is detected on the left or right side according to the principles of the present invention described above, the lane width thereby determined is fixed to its previous value and maintained over time until the end of the exit lane is detected. This makes it possible to maintain a reliable lane width. Next, based on the lateral position of the ground marking lines detected on the opposite side and the fixed value of the lane width, the ground marking lines located on the side where the exit lane is detected are reconstructed. In this way, the reconstructed ground marking lines follow the changes relative to the ground marking lines detected on the opposite side of the detected exit lane, but are transferred to the other side of the main lane, that is, to the position where the ground marking lines are located on the side where the exit lane is detected, on the basis of which it is desired to use the LCA system to control the trajectory of the vehicle. In other words, the correction applied to the ground marking lines located on the side where the exit lane is detected is applied by symmetric the ground marking lines detected on the opposite side with respect to the camera.

[0080] For example, according to Figure 3 For example, when the right coaxial exit lane EXIT_R is detected, at each time interval between the start and the end of the detection of the right coaxial exit lane EXIT_R, the position of the right ground marking line that delimits the main lane L on the side of the exit lane EXIT_R is corrected as follows:

[0081] [Math.4]

[0082] POS R，校正 (k) = pos L (k)+LW(k-1)

[0083] Among them, pos R,校正 (k) is the corrected lateral position of the right marker relative to the camera at time k,

[0084] POS L (k) is the lateral position of the left ground marker relative to the camera determined at time k, and

[0085] LW(k-1) is the lane width determined at time (k-1) before the right coaxial exit lane is detected.

[0086] This correction makes it possible to reconstruct the right ground markings that delimit the main lane L on the side where the right coaxial exit lane EXIT_R is detected. Figure 3 By L R_校正 express.

[0087] The corrected right ground marking line L R_校正 Symmetrically follow the left ground marker line L relative to the camera L This correction of the line located on the side of the detected exit lane thus allows the LCA system to have the correct line reference (i.e., LR_校正 ) to correct the vehicle's trajectory, thereby automatically keeping the vehicle in the center of its main lane L.

[0088] In the above Figure 4 , Figure 5 and Figure 6 In the case of the examples described in , this line correction principle is applied in the same way once an exit lane is detected. Thus, in all these cases, the ground markings on the side where the exit lane is detected are corrected so that the trajectory can be corrected to keep the vehicle in its main lane. For example, in Figure 5 In the case where the left standard exit lane is detected, the left standard exit lane EXIT_STD_L is actually detected by the first detection algorithm. According to the above principle, the corrected left ground marking line L is reconstructed. L_校正 , which follows the right ground marker line L symmetrically with respect to the camera R changes.

Claims

1. A method for detecting an exit lane for a motor vehicle (10) based on a series of images in front of the vehicle, the images being acquired by a camera on board the vehicle while the vehicle is moving along a main lane (L), the main lane being defined by a plurality of lanes along the right edge (L) of the lane. R ) and the left edge (L L ) is laterally delimited by a ground marking line located at a position, the method comprising: Detect the right ground marking line (L R ) and the left ground marking line (L L ), The step of determining a first heading angle of the vehicle relative to a detected right ground marking line and a second heading angle of the vehicle relative to a detected left ground marking line, A step of providing a first exit lane detection state according to a comparison between the value of the first heading angle or the value of the second heading angle and a predetermined threshold, wherein the first exit lane detection state represents that a right standard exit lane (EXIT_STD_R) or a left standard exit lane (EXIT_STD_L) is detected, respectively, The method is characterized in that the method further includes a step of providing a second exit lane detection state based on firstly comparing the derivatives of the changes of the first heading angle and the second heading angle with respect to time and secondly verifying that the value of the first heading angle or the value of the second heading angle respectively remains substantially close to a zero value, the second exit lane detection state representing detection of a right exit lane (EXIT_R) or a left exit lane (EXIT_L) each coaxial with the main lane (L).

2. The method according to claim 1, characterized in that The method includes arbitrating between the first exit lane detection state and the second exit lane detection state to grant processing priority to the first exit lane detection state when a second exit lane detection state has been provided.

3. The method according to claim 1 or 2, characterized in that When the time derivative of the second heading angle or the change of the first heading angle is respectively greater than the time derivative of the change of the first heading angle or the second heading angle plus a predefined minimum margin, and the value of the first heading angle or the value of the second heading angle is respectively zero to a predetermined tolerance threshold, the second exit lane detection status is provided for the right coaxial exit lane or the left coaxial exit lane, respectively.

4. The method according to claim 1 or 2, characterized in that: When the value of the first heading angle or the value of the second heading angle is respectively higher than the predetermined threshold, the first exit lane detection state is provided for the right standard exit lane or the left standard exit lane respectively.

5. The method according to claim 1, characterized in that The method includes the step of correcting a position of a ground marking line detected on a side of the exit lane detected according to the first exit lane detection state or the second exit lane detection state.

6. The method according to claim 5, characterized in that Correction of the ground marking line detected on one side of the detected exit lane includes reconstructing a corrected ground marking line on one side of the detected exit lane based on the position of the ground marking line detected on the opposite side relative to the detected exit lane and the measurement value of the lane width obtained using the camera.

7. The method according to claim 6, characterized in that The measured value of the lane width includes a measured value fixed to a measured value taken immediately before the exit lane is detected.

8. The method according to claim 6 or 7, characterized in that The method comprises the step of correcting the lateral position of the vehicle in its lane based on the position of the detected ground marking lines on the opposite side relative to the detected exit lane and the position of the corrected ground marking lines on the side of the detected exit lane.

9. A device for detecting an exit lane for a motor vehicle traveling on a main lane laterally delimited by ground markings, the device being characterized in that the device comprises at least one processing unit adapted to cooperate with a front camera equipped on the vehicle and to implement the steps of the method as described in any one of claims 1 to 8.

10. A motor vehicle, characterized in that: The motor vehicle comprises at least one on-board front camera and the device as claimed in claim 9.

Images