摘要
针对车辆在行驶时摄像机所拍摄图像的晃动问题,包括平移以及旋转抖动,文章提出了一种新型的以高速公路车道线以及其交叉点构成的三角形作为图像特征的数字稳像方法。首先依据车载视频图像中车道线的分布特征,对待处理图像进行分割,以及图像边缘信息进行加强,然后再进行图像特征的提取构造。有效提高了算法的精度并且减少了计算量。实验表明,基于车道线特征的数字稳像算法具有更高的稳定性,并且很少受到天气、过往车辆等外部因素的影响。
According to the vehicle in the road taken camera image sloshing problems, including translation and rotation jitter, present a digital image stabilization method for mobile video surveillance based on a lane line and its intersection points of a triangle formed. Ac cording to lane distribution of the image, make image segmentation first, intensify the edge information of image, then extract and con struct the image feature. It can effectively enhance the accuracy and reduce the computation. The experiment proves that this method is more effective and rarely affected by weather, passing vehicles and other external factors.
出处
《计算机技术与发展》
2013年第3期1-4,共4页
Computer Technology and Development
基金
国家"863"高技术发展计划项目(2012AA011804)
关键词
数字稳像
HOUGH变换
特征匹配
图像分割
运动估计
运动补偿
digital image stabilization
hough transform
feature matching
image segmentation
motion estimation
motion compensation
