摘要
为了提高特征点匹配的准确率,本文提出了一种基于改进混合滤波、特征描述符降维、SIFT特征匹配、RANSAC剔除误匹配点以及PSO算法的特征点匹配。首先将场景图像进行滤波处理达到去噪效果,然后通过特征描述符降维以减少计算量,再通过RANSAC对基于SIFT的特征点匹配进行误匹配的剔除,最后使用PSO算法进行优化以寻找到最佳的Ratio值。通过在模糊、较暗、较亮和遮挡4种以机械手为背景的场景下的图像,进行4种算法的对比实验,最后表明本文算法的误匹配率最小,精确度最高。
In order to improve the accuracy of feature point matching,this paper proposes a feature point matching based on improved hybrid filtering,feature descriptor dimensionality reduction,SIFT feature matching,RANSAC eliminating false matching points and PSO algorithm.Firstly the scene image is filtered to achieve denoising effect,then the feature descriptor is used to reduce the dimensionality to reduce the amount of calculation,and then the SIFT-based feature point matching is used to eliminate the false matching through RANSAC.Finally,the PSO algorithm is used for optimization to find the most optimal Ratio value.Through the comparison experiments of the four algorithms in the blurred,darker,brighter and occluded scenes with the manipulator as the background,the results show that the algorithm in this paper has the smallest mismatch rate and the highest accuracy.
作者
徐澳
华云松
夏春蕾
陈诗雨
XU Ao;HUA Yunsong;XIA Chunlei;CHEN Shiyu(School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093;College of Ocean Science and Engineering,Shanghai Maritime University,Shanghai 201306)
出处
《软件》
2022年第9期83-86,119,共5页
Software
作者简介
徐澳(1996-),男,江苏徐州人,硕士研究生,研究方向:图像处理、机器视觉。
