基于仿射不变约束与快速EKF滤波的航拍图像稳像

考虑到航拍机载成像平台抖动严重、视频稳像匹配环节精度不一致的特点以及航拍图像稳像技术快速、准确的要求,提出了一种结合仿射不变约束与快速扩展卡尔曼(Extend Kalman Filter,EKF)滤波的图像稳像算法。该算法首先以视频参考帧中的角点量作为特征点,通过Harris检测器选择出稳定角点；然后对待配准点构建Delaunay三角网进行初始匹配,提出利用仿射不变约束方法筛选出精确匹配点；最后利用快速EKF运动滤波方法实时估计和修正噪声的统计特性,从而解决摄像机扫描运动中存在的抖动问题。在对大量分辨率为640×480pixel的航拍图像的仿真实验中,可通过仿射不变约束实现精确的模型估计,采用的快速运动补偿方法在补偿过程中耗时为5.054ms,比传统的运动补偿方法节约了69.5%的时间。实验结果表明,该算法能够实时稳定航拍视频帧间的抖动现象,并能有效跟随场景的真实扫描。

Aerial Video Image Stabilization Based on Affine Invariant Constraint and Fast EKF Adaptive Filter

In view of serious jitter of aerial airborne imaging,different accuracies of obtained matching points in aerial video stabilization,and requirements of fast and precise aerial image stabilization technology,an image stabilization algorithm was proposed combining affine invariant constraint and Extend Kalman Filter (EKF) adaptive filter.Firstly,corners are detected from reference frame as feature points,the locally stable points are selected by the Harris detector.Then Delaunay triangulation is used to find initial matching,and the most accurate matching points which best satisfy the affine invariant constraint are filtered.Finally,EKF motion filter method is used to estimate and correct the statistical property of noise,so jitter of the camera can be removed but scanning motion can be retained simultaneously.As to a large number of simulation experiment of aerial images with a resolution of 640 pixel ×480 pixel,accurate model can be estimated by affine invariant constraints,the rapid motion compensation method takes 5.054 ms in the process of compensation,and saves for 69.5% than traditional motion compensation method.The experimental results illustrate that the proposed algorithm can stabilize the inter-frame jitter and track the real scene effectively.