基于光流分析的车辆转弯及上下坡信息获取

传统的前照灯系统存在安全隐患,自适应前照灯系统(AFS)可以有效消除安全隐患、降低交通事故率,准确获取行车环境特征信息是AFS实现自适应照明的前提条件.文章采用摄像头获取车辆夜间行车时驾驶员视野前方视频图像序列,提出了一种图像区域分割方法,结合基于光流模型的图像运动估计方法计算出各个区域的所有像素点的光流速度,并依据不同区域光流速度提出车辆行车环境特征判断准则,来获取夜间行车转弯及上下坡信息,实验证明了该方法的有效性.

Road environment information acquisition for vehicles based on optical flow analysis

The traditional headlight system has some hidden problems of safety, but the adaptive front-lighting system(AFS) can eliminate these problems and reduce the rate of traffic accidents effectively. Achieving vehicle running surrounding feature information accurately is the precondition of AFS implementation. In this paper,the image sequence is obtained from the video data of driver's vision when the vehicle is running at night. A kind of image region segmentation method is proposed and the image motion estimation method based on the optical flow model is used to calculate the optical flow velocity of each pixel in four regions, then a rule of judging vehicle running surrounding features is proposed according to the velocity to obtain road environment information when the vehicle is running at night. The effectiveness of the method is proved.