摘要
障碍检测是基于车载视觉感知系统的车辆自主避撞技术的基础。针对这一问题提出了一种障碍目标检测算法。首先,为了提高道路边缘检测的效率,采用区域Hough变换完成道路边缘检测,进而提取道路区域;然后,在提取出的道路区域内利用图像分割算法完成障碍目标的提取;由于分割出的二值图像中目标表现为独立的大小不一的区域,为此给出标识归属方法辨别目标区域。最后,利用自适应区域生长算法得到完整的障碍目标。相关实验结果证明了算法是可行的。
Obstacle detection is the basic function of Independent collision avoidance based on vehicle-mounted visual system. A relative algorithm is proposed. It can be divided into three parts. Firstly, in order to improve the efficiency of the road edge detection, the detected image is divided into many regions of different sizes, and in ever- y region Hough Transformation is used for road edge detection respectively. Second, the targets region is separated from the extracted pure road region with image segmentation method. Because target obtained in segmentation are discrete regions of varying sizes in the segmented binary image, Target identification method is presented to distinguish the discrete regions. At last, adaptive regional growth method is employed to get the whole target in the last part of this paper. Experimental results show that the algorithm is feasible.
出处
《科学技术与工程》
北大核心
2013年第15期4449-4453,共5页
Science Technology and Engineering
基金
中国民航科技项目(MHRD201124)
国家自然科学基金(61203170)
中央高校基本科研业务费(NS20122026)资助
关键词
目标检测
HOUGH变换
阈值分割
区域生长
target detection Hough transform threshold segmentation regional growth
