| 摘要 |
The invention relates to a method (200) for identifying a position of a vehicle (100) in a lane (320). The method comprises four steps (210, 220, 230, 240). In the read-in step (210), a first image of a first camera (120b) with a first optical axis (130b) and an at least second image of an at least second camera (120a) with a second optical axis (130a) are read in, wherein the first (120b) and second (120a) camera are oriented in such a way that the first optical axis (130b) and the second optical axis (130a) intersect outside the detection angle (135b) of the first camera (120b) and/or of the second camera (120a), and wherein the first image and the at least second image each form a viewing angle which deviates at most by an angle of 90 degrees from the direction (150) of travel of the vehicle (100). In the determination step (220), at least one interlinked image gradient (610) is determined in the first image and at least one further interlinked image gradient (610) is determined in the at least second image. In the assignment step (230), the at least one interlinked image gradient (610) is assigned to a carriageway boundary object (330, 335, 340, 350, 410) and the at least one further interlinked image gradient (610) is assigned to the carriageway boundary object (330, 335, 340, 350, 410) and/or to a further carriageway boundary object (330, 335, 340, 350, 410), and in the calculation step (240) a relative position of the vehicle (100) in relation to the carriageway boundary object (330, 335, 340, 350, 410) and/or to the further carriageway boundary object (330, 335, 340, 350, 410) is calculated in order to identify the position of the vehicle (100) in the lane (320). |
