| 摘要 |
An oncoming vehicle is detected on the basis of images captured by CCD cameras when an own vehicle is making a turn with passing over an oncoming lane at an intersection. On this occasion, after the own vehicle has reached a predetermined turning state, by extracting three-dimensional objects having speed components Vnz not lower than a first threshold in the longitudinal direction of the own vehicle and in a direction along which the oncoming vehicle is coming closer to the own vehicle and having other speed components Vnx not lower than a second threshold in the lateral direction of the same and by detecting an oncoming vehicle out of the extracted three-dimensional objects, even in the case, for example, where the oncoming vehicle apparently moves in an image screen from obliquely left ahead of the own vehicle toward obliquely right behind of the same, the oncoming vehicle is properly detected.
|
