摘要
从图像恢复的迭代处理原理出发,针对简单的迭代方法中诸如每次的恢复误差在迭代过程中不断反馈与积累从而造成恢复图像的严重失真等问题,提出了一种新的迭代处理方法.新方法从每次迭代的恢复结果中提取新的信息,然后将该新信息和以前的恢复结果进行信息传递,并在此基础上设置了一个用于调节最终迭代结果的平滑程度的参数,使新方法具有更强的灵活性.新方法对真实SAR图像的处理结果表明该方法能有效的克服简单迭代方法中的误差积累以及迭代结果不收敛等不足.
According to the principle of iterated processing of images' restoration, and aiming at some problems in the simply iterated methods such that errors' feedback and accumulation make restored imagesserious distortion, this paper provides a new iterated method. The new method extracts the new information from the results of every iteration, and then transfers the new information to the previously restored results, additionally sets up a parameter to adjust the smoothing degree of the final result based on the information transfer. The parameter makes the new method more flexible. The results of test on the real SAR image shows that the new method can overcome some defects of simply iterated method such as errorsaccumulation, final result's unconvergence and so on.
出处
《武汉大学学报(理学版)》
CAS
CSCD
北大核心
2002年第3期355-359,共5页
Journal of Wuhan University:Natural Science Edition
基金
国家自然科学基金(60072041)
国防科技重点实验室基金资助项目