一种实用的小目标配准方法

Applied method for small target registration

图像配准是图像融合技术的基本环节和首要问题,只有经过配准后的图像才能进行有效的融合。其中,小目标由于几乎无特征信息可以利用,所以常规的配准方法都不适用。针对图像识别中小目标的配准问题,分析了其配准特点,创新性地提出了先配准目标视场,再配准目标位置的方法,提出了视场配准的概念。首先运用成像原理,用焦距、分辨率和像元尺寸建立不同CCD之间的视场对应关系,利用此关系完成目标视场的截取放大,使不同CCD得到的图像视场一样大。然后在分析通常采用的最小平均绝对误差(MAD)相关匹配方法缺陷的基础上,提出用最多近邻点距离(MCD)的匹配方法来对准目标位置,完成目标质心的配准。实验结果表明,此方法可以很好地配准小目标,且误差不超过2个像素。由于其针对性强,因而具有较强的实际应用价值。

The registration of multi-sensor images is a basic and key technique in image fusion processing,image fusion can be done effectively only after registration.Normal registration schemes are not suitable for small target because there are nearly no features can be used.Aiming to the characteristics of small target registration,a new method of completed the registration by two steps is presented,and the idea of FOV registration is brought forward.Firstly,using imaging principle,with focus,resolution and pixel dimension of CCD,the FOV corresponding relation between different CCD is established,and then it is used to complete FOV registration by cutting and zooming out the FOV of target. Secondly,based on analyzing the disadvantage of the common used Minimum Absolute Deviation（MAD）matching,the method using Maximum Close Distance （MCD）matching to register target position is proposed.The experimental result shows that the method can register small target well,and the error is not beyond two pixels.Because it＇s strong pertinence,it can be used in project.