一种遥感图象自动配准方法
|
摘要
遥感图象的配准是多项重要的图象处理工作(如图象融合、图象比较、图象计算等)的前提。该文提出了一种简单可靠的遥感图象自动配准方法,该方法利用待配准图象之间的灰度相关性,首先采用改进的有利算子(兴趣算子)自动地在基准图象上搜寻特征点,然后进行特征点匹配,在待配准图象上搜寻特征点的同名点,并按相关性的大小从中筛选出有效的控制点,构成不规则三角网(TIN),最后对三角网中的每一个三角形进行多项式拟合和插值变换。与传统的人工输入控制点方法相比,该文提出的方法具有更高的精确度,而且节省了人力和时间,实验证明具有一定的实用性。
The registration of remote sensing image is the premise of many important image processes(e.g.image fusion,compare and calculation).In this paper,a simple and reliable method of automatic remote sensing image registration is proposed,which is based on the correlation of the gray level of the images.First,this paper uses the modified interest operator to search the feature points on the normalized image.Second,it matches these feature points to find their homologous points on the source image,and then selects the effectual control points according to the value of correlation,then makes them into Triangular Irregular Network(TIN).At last it translates every triangle of the TIN by polynomial fitting and insertion algorithm.Compared with the traditional method for registration which needs unnatural control points,the algorithm has the following advantages:(1)It has higher accuracy.(2)It economizes manpower and time.(3)The experiment results show that this algorithm is simple and useful.
The registration of remote sensing image is the premise of many important image processes(e.g.image fusion,compare and calculation).In this paper,a simple and reliable method of automatic remote sensing image registration is proposed,which is based on the correlation of the gray level of the images.First,this paper uses the modified interest operator to search the feature points on the normalized image.Second,it matches these feature points to find their homologous points on the source image,and then selects the effectual control points according to the value of correlation,then makes them into Triangular Irregular Network(TIN).At last it translates every triangle of the TIN by polynomial fitting and insertion algorithm.Compared with the traditional method for registration which needs unnatural control points,the algorithm has the following advantages:(1)It has higher accuracy.(2)It economizes manpower and time.(3)The experiment results show that this algorithm is simple and useful.
引文
1.Morave'cH.VisualMappingbyaRobotRover犤C犦.In:Proc6thIntJointConfArtificialIntellTokyo,Japan,1979-08;1
2.FrstnerW.AFeaturebasedCorrespondenceAlgorithmforImageMatching.IntArchofPhotogrammetry,Rovaniemi,Finland,1986;26-III
3.吕言.用于提取数字图像点特征之有利算子研究犤J犦.测绘学报,1989;18(3):175～182
4.HannahMJ.ComputerMatchingofAreasinStereoImagery犤D犦.PhDdissertation.AIM239,ComputerScienceDepartment,StanfordUniver-sity,StanfordCalifornia,USA,1974
5.LuhmannT,AltroggeG.InterestOperatorforImageMatching犤C犦.In:IntercommissionConferenceonFastProcessingofPhotogrammetricData,Interlaken,Switzerland,1987-06
6.张永生.遥感图象信息系统犤M犦.北京:科学出版社,2000:87～90
7.宁书年等.遥感图象处理与应用犤J犦.北京:地震出版社,1995
8.武晓波,王世新,肖春生.一种生成Delaunay三角网的合成算法犤J犦.遥感学报,2000;4(1):32～35
2.FrstnerW.AFeaturebasedCorrespondenceAlgorithmforImageMatching.IntArchofPhotogrammetry,Rovaniemi,Finland,1986;26-III
3.吕言.用于提取数字图像点特征之有利算子研究犤J犦.测绘学报,1989;18(3):175～182
4.HannahMJ.ComputerMatchingofAreasinStereoImagery犤D犦.PhDdissertation.AIM239,ComputerScienceDepartment,StanfordUniver-sity,StanfordCalifornia,USA,1974
5.LuhmannT,AltroggeG.InterestOperatorforImageMatching犤C犦.In:IntercommissionConferenceonFastProcessingofPhotogrammetricData,Interlaken,Switzerland,1987-06
6.张永生.遥感图象信息系统犤M犦.北京:科学出版社,2000:87～90
7.宁书年等.遥感图象处理与应用犤J犦.北京:地震出版社,1995
8.武晓波,王世新,肖春生.一种生成Delaunay三角网的合成算法犤J犦.遥感学报,2000;4(1):32～35
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心