一种基于双模式成像的遥感图像去运动模糊方法  

Motion Blur Suppression Method Based on Dual Mode Imaging for Remote Sensing Image

作　　者：李奇[1] 董文德[1] 徐之海[1] 冯华君[1] 陈跃庭[1] 

机构地区：[1]浙江大学现代光学仪器国家重点实验室,杭州310027

出　　处：《航天返回与遥感》2013年第4期86-92,共7页Spacecraft Recovery & Remote Sensing

摘　　要：对地观测载荷大多工作在颤振环境之中,这种颤振造成光轴偏移、像面抖动,由此产生的运动模糊对高分辨率成像与高精度探测产生了极为不利的影响。对空间相机像面去模糊的工作,大致分为两种方案:一种为硬件实时探测加硬件实时补偿,另一种为硬件实时探测加事后软件补偿。文章提出通过主成像相机的双成像模式获得低信噪比无抖动图像和正常信噪比模糊图像,利用图像处理手段从这一对图像中求得抖动信息(模糊核),并以此为先验知识指导模糊图像的复原,抑制像面抖动。这种方法既不需要补偿器件,也不需要探测元件和探测支路,并在图像迭代处理过程中不断修正模糊核,在初始模糊核有一定误差的情况下,同样可望取得良好的像面去模糊效果,对于遥感图像像质提高具有潜在的应用价值。Most space optical imaging systems work in vibration environment, so optical axis motion and image plane jitter caused by the vibration result in motion blur of image, which is a serious impediment to high resolution imaging and high precision detection. Image deblur of space camera has two schemes to achieve mo- tion blur suppression： the first is combination of real-time hardware detection and real-time hardware compen- sation, and the second is combination of real-time hardware detection and software restoration compensation. The paper proposes a dual mode imaging of space camera to obtain two images： the first image is an inadequate exposure motion blur-free image and the second image is a precise exposure blurred image. Using image proc- essing tools, we can obtain motion path during exposure （blur kernel） from the two images, and restore the sharp image with the blur kemel as a priori knowledge. This method does not require compensation devices, detection devices and auxiliary detection optical path, and can automatically revise blur kernel during iterative processing. The method can be easily applied in the current space camera for remote sensing and has potential application value to improve remote sensing image quality.

关 键 词：运动模糊 双模式成像 模糊核 图像复原 运动探测 空间相机 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]