基于B-SURF算法的图像运动估计
|
摘要
序列图像的运动矢量估计在电子稳像系统及图像超分辨率重建过程中都非常重要,传统的运动矢量估计算法大多需要提取相邻序列图像特征,根据相邻序列图像的特征点进行运动矢量估计,耗时相对较多,且难以达到实时性的要求.故提出并实现了一种基于SURF特征提取算法和BRIEF特征描述算子的运动估计方法:B-SURF方法,使用SURF特征提取算法思想对图像进行特征提取,采用BRIEF特征描述算子对检测到的特征点进行特征描述.由于引入了二进制编码,极大的降低了特征描述和匹配的计算量,提高了运动估计的计算速度,使算法能满足实时性需求.实验表明,实现方法在运算速度上有明显提升.
Motion vector estimation for sequence images plays an important role in electronic image stabilization system and in the image super-resolution reconstruction process.The traditional motion vector estimation algorithm often has to extract the image characteristics of the adjacent sequences to estimate the motion vector according to the feature point of the adjacent sequence images,which is relatively time-consuming and difficult to achieve the real-time requirement.Therefore a motion estimation algorithm based on SURF feature extraction algorithm and BRIEF feature description operator is put forth:the B-SURF method,which adopts SURF algorithm for image feature extraction,and then uses the BRIEF description operator for character description of feature points that are detected.Owing to the introduction of binary coding,it drastically minimizes the amount of calculation of the character description and matching,and thus enhances the computation speed of motion estimation and makes the algorithm meet the real-time requirement.Experiments show that the method proposed in this paper has a significant improvement in the speed of computational operation.
Motion vector estimation for sequence images plays an important role in electronic image stabilization system and in the image super-resolution reconstruction process.The traditional motion vector estimation algorithm often has to extract the image characteristics of the adjacent sequences to estimate the motion vector according to the feature point of the adjacent sequence images,which is relatively time-consuming and difficult to achieve the real-time requirement.Therefore a motion estimation algorithm based on SURF feature extraction algorithm and BRIEF feature description operator is put forth:the B-SURF method,which adopts SURF algorithm for image feature extraction,and then uses the BRIEF description operator for character description of feature points that are detected.Owing to the introduction of binary coding,it drastically minimizes the amount of calculation of the character description and matching,and thus enhances the computation speed of motion estimation and makes the algorithm meet the real-time requirement.Experiments show that the method proposed in this paper has a significant improvement in the speed of computational operation.
引文
[1]赵亚湘,刘少强,樊晓平.结合运动矢量和像素递归的全局运动估计方法[J].中国图象图形学报,2018,17(2):191-196.
[2]PUGLISI G A,BRUNA A R,BATTIATO S A.RandomTemporal Block Selection for Video Stabilization[C].California:Proceedings of IS&T/SPIE Electronic Imaging.2011.
[3]PUGLISI S B.Fast Block Based Local Motion Estimation for Video Stabilization[C].America:Computer Vision&Pattern Recognition Workshops.IEEE,2012.
[4]TIEYAN L,XUDONG Z,DESHENG W.An Improved Three-Step Search Algorithm For Block Motion Estimation[C].Shanghai:Conference on Communications.IEEE,2002.
[5]TANG J,HAN X,YUAN Z,et al.An Approach of Electronic Image Stabilization Based on the Representative Point Matching[C].Yangon:International Conference on Genetic&Evolutionary Computing.IEEE,2009.
[6]ZHAO H,WANG T.Image stabilization algorithm based on multi-bitplane[C].City:International Symposium on Photoelectronic Detection&Imaging:Advances in Imaging Detectors&Applications.International Society for Optics and Photonics,2009.
[7]刘玉峰,李震.基于改进位平面匹配法的运动估计[J].中国传媒大学学报(自然科学版),2014,21(6):48-55.
[8]王彦锟,刘方.采用位平面和钻石搜索的全局运动估计[J].兵工自动化,2007,26(11):80-82.
[9]姜思宇.基于SIFT的全局运动估计算法研究[D].秦皇岛:燕山大学,2011.
[10]TIAN S,ZHAO P,WANG N,et al.Aims at moving objects'improvement based on gray projection of algorithm of the electronic image stabilization[C].Beijing:International Congress on Image&Signal Processing.IEEE,2010.
[11]吉淑娇,朱明,雷艳敏,等.基于改进运动矢量估计法的视频稳像[J].光学精密工程,2015,23(5):1458-1466.
[12]曹雪情,戴越越,陈瑞,等.基于半像素运动估计的DCVS残差重建算法[J].计算机应用研究,2017,34(6):1867-1870.
[13]余应淮,谢仕义,梅其祥.基于核回归修正的上采样相位相关精确运动估计算法[J].计算机应用,2016,36(8):2316-2321.
[14]姚睿,郭庆新,孙艳梅,等.面向H.264的多级自适应快速整像素运动估计算法[J].吉林大学学报(工学版),2016,46(1):318-326.
[15]BAY H.SURF:Speeded up robust features[C].Graz:European Conference on Computer Vision,2006.
[16]杨海燕,罗文超,刘国栋.基于SURF算法和SC-RANSAC算法的图像配准[J].计算机应用研究,2013,30(5):1586-1588.
[17]CALONDER M,LEPETIT V,STRECHA C,et al.BRIEF:Binary Robust Independent Elementary Features[C].Heraklion:European Conference on Computer Vision,2010.
[18]周莉莉,姜枫.基于FAST和BRIEF的图像匹配算法[J].计算机工程与设计,2015,36(5):1269-1273.
[2]PUGLISI G A,BRUNA A R,BATTIATO S A.RandomTemporal Block Selection for Video Stabilization[C].California:Proceedings of IS&T/SPIE Electronic Imaging.2011.
[3]PUGLISI S B.Fast Block Based Local Motion Estimation for Video Stabilization[C].America:Computer Vision&Pattern Recognition Workshops.IEEE,2012.
[4]TIEYAN L,XUDONG Z,DESHENG W.An Improved Three-Step Search Algorithm For Block Motion Estimation[C].Shanghai:Conference on Communications.IEEE,2002.
[5]TANG J,HAN X,YUAN Z,et al.An Approach of Electronic Image Stabilization Based on the Representative Point Matching[C].Yangon:International Conference on Genetic&Evolutionary Computing.IEEE,2009.
[6]ZHAO H,WANG T.Image stabilization algorithm based on multi-bitplane[C].City:International Symposium on Photoelectronic Detection&Imaging:Advances in Imaging Detectors&Applications.International Society for Optics and Photonics,2009.
[7]刘玉峰,李震.基于改进位平面匹配法的运动估计[J].中国传媒大学学报(自然科学版),2014,21(6):48-55.
[8]王彦锟,刘方.采用位平面和钻石搜索的全局运动估计[J].兵工自动化,2007,26(11):80-82.
[9]姜思宇.基于SIFT的全局运动估计算法研究[D].秦皇岛:燕山大学,2011.
[10]TIAN S,ZHAO P,WANG N,et al.Aims at moving objects'improvement based on gray projection of algorithm of the electronic image stabilization[C].Beijing:International Congress on Image&Signal Processing.IEEE,2010.
[11]吉淑娇,朱明,雷艳敏,等.基于改进运动矢量估计法的视频稳像[J].光学精密工程,2015,23(5):1458-1466.
[12]曹雪情,戴越越,陈瑞,等.基于半像素运动估计的DCVS残差重建算法[J].计算机应用研究,2017,34(6):1867-1870.
[13]余应淮,谢仕义,梅其祥.基于核回归修正的上采样相位相关精确运动估计算法[J].计算机应用,2016,36(8):2316-2321.
[14]姚睿,郭庆新,孙艳梅,等.面向H.264的多级自适应快速整像素运动估计算法[J].吉林大学学报(工学版),2016,46(1):318-326.
[15]BAY H.SURF:Speeded up robust features[C].Graz:European Conference on Computer Vision,2006.
[16]杨海燕,罗文超,刘国栋.基于SURF算法和SC-RANSAC算法的图像配准[J].计算机应用研究,2013,30(5):1586-1588.
[17]CALONDER M,LEPETIT V,STRECHA C,et al.BRIEF:Binary Robust Independent Elementary Features[C].Heraklion:European Conference on Computer Vision,2010.
[18]周莉莉,姜枫.基于FAST和BRIEF的图像匹配算法[J].计算机工程与设计,2015,36(5):1269-1273.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |