摘要
针对图像特征在匹配过程中产生误匹配、漏匹配的问题,对SURF算法与误匹配剔除算法进行了相关的研究.提出最小距离与随机抽样一致性(RANSAC)两种方法相融合的误匹配消除算法.算法首先通过设置适当阈值的最小距离法对特征之间的匹配进行粗略的淘汰,然后利用优化了的RANSAC对其进行精确的剔除.最后,从数据集中的随机选取两帧图像对算法进行了实验验证,表明提出的算法提高了误匹配剔除的效率的同时,保持了正确匹配的数量.
Aiming at the problem of mismatching and missing matching in the matching process of image features,SURF algorithm and mismatch elimination algorithm were studied. On this basis,a mismatch elimination algorithm based on the two methods of minimum distance and random sampling consistency( RANSAC) was proposed. First,the algorithm was used to eliminate the matching between features by setting the minimum distance method of the appropriate threshold,and then the optimized RANSAC was used to eliminate it accurately. Finally,the algorithm was tested through random selection of two frame images from the data set. It shows that the algorithm proposed in this paper improves the efficiency of mismatch rejection and keeps the number of correct matching.
作者
李文彬
刘璇
张建畅
张建华
LI Wen-bin;LIU Xuan;ZHANG Jian-chang;ZHANG Jian-hua(School of Mechanical Engineering,Hebei University of Technology,Tianjin 300130,China)
出处
《计算机仿真》
北大核心
2019年第10期233-237,共5页
Computer Simulation
基金
河北省杰出青年科学基金(F2017202062)