摘要
针对深空目标与空间相机相对运动造成的图像运动模糊问题,文章提出了一种基于图像形态学的运动模糊复原方法。首先运用图像形态学细化的方法消除图像中的冗余信息,快速提取图像中目标的特征;其次运用最小二乘法直线拟合方法对目标运动方向进行估计,再利用连通域划分结果对目标运动长度进行估计;最后采用约束最小二乘方滤波对退化模型进行逆推导运算来实现目标运动模糊复原。通过实验验证结果表明:文章提出的基于图像形态学的深空图像运动模糊复原方法能够正确估计运动模型参数,通用性强,易于在轨实现,可以处理目标与相机相对运动造成的运动模糊问题,在目标运动方向不一致的图像中取得了良好的复原效果。
Aiming at the problem of image motion blur caused by relative motion between deep space target and space camera,this paper proposes a motion blur restoration method based on image morphology.Firstly,the method of image morphological refinement is used to eliminate the redundant information in the image,and the characteristics of the target in the image are quickly extracted.Secondly,the least squares line fitting method is used to estimate the movement direction of the target,and then the connected domain division results are used to estimate the movement of the target.Finally,the constrained least squares filter is used to inversely derive the degradation model to achieve the target motion blur restoration.The experimental verification results show that the motion blur restoration method for deep space images based on image morphology proposed in this paper can correctly estimate the parameters of the motion model.Good restoration results are achieved in images with inconsistent motion directions of multiple objects.
作者
南诺
侯作勋
曹东晶
周海岸
唐琦
李馨
NAN Nuo;HOU Zuoxun;CAO Dongjing;ZHOU Haian;TANG Qi;LI Xin(Beijing Institute of Space Mechanics&Electricity,Beijing 100094,China)
出处
《航天返回与遥感》
CSCD
北大核心
2023年第2期101-108,共8页
Spacecraft Recovery & Remote Sensing
关键词
拖尾
运动模糊
滤波复原
图像处理
航天遥感
trailing
motion blur
filter recovery
image processing
space remote sensing
