摘要
研究了基于计算机视觉技术的高速公路上的障碍物实时检测问题。其目的是为了得到障碍物的准确图象位置信息 ,实现汽车自动防撞安全行驶。基于图象分割、模式识别理论 ,笔者提出了一种根据汽车整体模型并结合高宽值来识别高速公路上主要障碍物———汽车的检测方法。而且利用有色噪声的卡尔曼滤波器实现了运动物体的实时跟踪。
This paper studies the problem of computer vision in the vehicle intelligent auxiliary driving system. An effective and real time method in highway to search and track obstacle is presented. Obstacle candidates are tracked and analyzed, false alarms generated by the object detector are recognized and rejected. It also supplies some valuable discussion on problems existing in its experiments.
出处
《重庆大学学报(自然科学版)》
CAS
CSCD
2000年第4期123-126,137,共5页
Journal of Chongqing University
基金
国家自然科学基金资助项目 !(696740 1 2 )
关键词
计算机视觉
透视投影
汽车操作
障碍物
computer vision
linear projection
image segment
pattern recognition
Kalman filtering