[1]吴强,朱庆保,沈玲玲,等.基于仿射光流场的视频图像镶嵌技术[J].南京师范大学学报(工程技术版),2007,07(04):072-75.
 Wu Qiang,Zhu Qingbao,Sheng Lingling.Video Mosaicing Based on Affine Optical Flow Field[J].Journal of Nanjing Normal University(Engineering and Technology),2007,07(04):072-75.
点击复制

基于仿射光流场的视频图像镶嵌技术
分享到:

南京师范大学学报(工程技术版)[ISSN:1006-6977/CN:61-1281/TN]

卷:
07卷
期数:
2007年04期
页码:
072-75
栏目:
出版日期:
2007-12-30

文章信息/Info

Title:
Video Mosaicing Based on Affine Optical Flow Field
作者:
吴强;朱庆保;沈玲玲;
南京师范大学数学与计算机科学学院, 江苏南京210097
Author(s):
Wu QiangZhu QingbaoSheng Lingling
School of Mathematics and Computer Science,Nanjing Normal University,Nanjing 210097,China
关键词:
视频图像镶嵌 仿射光流 图像配准
Keywords:
v ideo mosa ic ing a ffine op tica l flow field im age reg iste ring
分类号:
TP391.41
摘要:
为获取宽视场的场景图像,提出了一种利用仿射光流场进行视频图像镶嵌的方法,首先根据光流约束方程,建立了关于图像间仿射变换参数的目标函数,然后通过最小化目标函数计算出图像间的仿射变换参数,再根据计算出来的仿射变换参数实现图像的配准,最后通过图像融合获取无缝镶嵌场景图像.实验表明,该方法可以有效地避免误拼接,可以获得较好的镶嵌结果.
Abstract:
To obtain a scene im ag e o f w ide v iew, a v ideo m osa ic ing m ethod based on affine op tica l flow fie ld is put forward. The object function on the a ffine transform param e ters betw een v ideo fram es is bu ilt acco rd ing to optica l constra int equa tion. Then the a ffine transform param eters a re ca lculated bym in im iz ing the object func tion. Acco rd ing to the ca lcu la ted a ffine transform param ete rs, reg ister ing im ages are rea lized. Im age b lend ing is pu t up to im plem ent video mo sa ic ing w ith these affine transfo rm param eters. Expe rim en tal results using v ideo sequence w ith rea l scene have demonstra ted that the m ethod can e ffective ly avo id inaccu rate m osa icing and obtain betterm osa icing im age.

参考文献/References:

[ 1] Iran iM, Anandan P, Berg en J, et a.l Effic ient representa tions of v ideo sequences and the ir app lications[ J]. Signa l Processing: Im age Communication, 1996 ( 8) : 327- 351.
[ 2] 阮宗才, 梁栋, 吴福朝, 等. 一种自动全景拼图的快速算法[ J]. 计算机研究与发展, 2001, 38( 5): 536- 539.
Ruan Zongca,i L iang Dong,W u Fuchao, e t a.l A fast algorithm for autom ated pano ram ic mosaic ing [ J]. Journa l of Com puter Research & Deve lopm ent, 2001, 38( 5) : 536- 539. ( in Ch inese)
[ 3] H su Chiouting, Cheng T zuhung, Beuker, et a.l Fea ture based v ideo mo sa ic [ C] / / Proceed ing s of Interna tiona l Conference on Im ag e Processing. C anada: Vancouver, BC, 2000( 2): 887- 890.
[ 4] Venkatesw ar, Chellappa R. H ierarch ica l ste reo and m otion correspondence using featu re groupings[ J]. Internationa l Journa l o f Com puter V ision, 1995, 15: 245- 269.
[ 5] Sze lisk iR. V ideo m osa ics for v irtua l env ironm ents[ J]. IEEE Compu terG raph ics and Applications, 1996, 16( 2) : 22- 30.
[ 6] Y ihong G, Gu ido P, Dav id L. A robust im agem osa icing techn ique capable of creating integrated panoram as[ C] / / Proceed ing s o f IEEE Internationa l Con ference on Inform ation V isua lization. CA, USA: Los A lam itos, 1999: 24- 29.
[ 7] 郑南宁. 计算机视觉与模式识别[M ]. 北京: 国防工业出版社, 1998.
Zheng Nann ing. ComputerV ision and Pattern Recogn ization[M ]. Be ijing: TheN ationalDe fence Industry Press, 1998. ( in Ch-i nese)

备注/Memo

备注/Memo:
基金项目: 江苏省自然科学基金( kb2006218)资助项目.
作者简介: 吴 强( 1970-) , 讲师, 主要从事图像处理和智能控制等方面的教学与研究. E-m ail:wu qiang@ n jnu. edu. cn
更新日期/Last Update: 2013-04-29