A New Starry Images Motion Compensation Method Based on Delaunay Triangulation and ICP
-
摘要: 星图运动补偿技术是有效提高空间监测中复杂背景弱小目标检测精度的关键技术之一。本文中提出了一种基于Delaunay三角剖分和ICP算法相结合的星图运动补偿算法。该方法首先通过Delaunay三角剖分建立星图中恒星之间的线索矩阵,其次通过ICP配准算法得到相邻星图间的对应关系,并通过SVD最优解析得到相邻帧之间的变换关系即摄像机运动模型,最后通过双线性内插法进行运动补偿。实验结果表明:该方法可有效实现高精度的运动背景补偿,为复杂背景弱小目标检测奠定技术基础。
-
关键词:
- Delaunay三角剖分 /
- ICP算法 /
- 运动补偿
Abstract: Motion compensation method for starry image registration is one of the key technologies,which could effectively improve the dim target detection precision in complex background for space surveillance.In this paper,a new method combining with Delaunay triangulation and ICP algorithm is presented for starry images sequences motion compensation.Firstly,this method establishes the clue matrix for several fixed stars in one starry image by Delaunay triangulation.Secondly,this method could get the corresponding relations between the adjacent starry maps by using the ICP registration algorithm.And through the SVD optimal resolution to get the transform relation between the adjacent frames which is the camera motion model.Finally,the bilinear interpolation method is employed to finish the motion compensation.The experimental results show that,this method can effectively realize high precision motion background compensation,and lay the technical foundation for dim small target detection in complex background.-
Key words:
- delaunay triangulation /
- ICP algorithm /
- motion compensation
-
[1] Huang J C,Hesieh W S.Automatic feature-based global motionestimation in video sequences[J].IEEE Transactions on Con-sumer Electronics,2004,50(3):911~915 [2] 张春华,周晓中,王学伟,周晓东.一种面向弱小目标检测的序列星图配准算法[J].中国图象图形学报,2010,15(3):435~442 [3] 陈颖,李在铭.基于鲁棒回归的运动背景补偿技术[J].信号处理,2002,18(1):36~38 [4] Wang D M,Wang L M.Global motion parameters estimation usinga fast and robust algorithm[J].IEEE Transactions on Circuitsand Systems for Video Technology,1997,7(5):823~826 [5] Zhu Y,Hu W J,Zhou J,Duan F.A new starry images matchingmethod in dim and small space target detection[A],2009 FifthInternational Conference on Image and Graphics[C],2009,9:447~450 [6] BESL P J,M.N.D.A method for registration of 3D shapes[J].IEEE Transactions on Pattern Analysis and Machine Intelli-gence,1992,14(2):239~256 [7] 施友道,何曙光,傅琼波.一种改进的2D-Delaunay三角剖分算法[J].空军雷达学院学报,2010,24(2):137~139 [8] 孙瑾秋,周军,张臻.基于能量累加的空间目标星象质心定位方法[J].光学精密工程,2011,19(12):3043~3048 [9] Arun K S,Huang T S,B.S.D.Least squares fitting of two 3Dpoint sets[J].IEEE Transactions on Patter n Analysis andMachine Intelligence,1987,9(5):698~700
点击查看大图
计量
- 文章访问数: 182
- HTML全文浏览量: 20
- PDF下载量: 8
- 被引次数: 0