摘要
随机抽样一致性算法RANSAC(Random Sample Consensus)是影像匹配领域中应用最广泛的鲁棒性算法之一。RANSAC算法能很好的解决误匹配问题,但是它的效率较低。本文提出一种新的影像匹配算法:算法在初始模型估计过程中,采用六点算法,降低算法迭代次数和抽样时间,提高了RANSAC算法的整体效率;并在数据检验过程中,定义一个误差加权函数,提高了算法的鲁棒性。实验结果表明,此算法提高了RANSAC算法的鲁棒性,计算速度高于RANSAC算法。
RANSAC algorithm is one of the most widely used robust algorithm in the field of image matching.RANSAC algorithmsolve the problem of false matches very well,but,it＇s efficiency is low.The paper gives a new algorithm for image matching：In theprocess of evaluating the initial model,six-points algorithm is used.It can reduce the number of the algorithm ＇s iterations and sampling time.The new image matching algorithm also define an error weighted function,it promote RANSAC＇s robust.The new algorithm is evaluated on real-world images,a significant increase in speed is shown,and the robust is better than RANSAC.
出处
《微计算机信息》
2010年第29期12-14,共3页
Control & Automation
