“北京一号”多光谱影像复原方法研究及实现

Research and Implementation of Multispectral Image Restoration of Microsatellite Beijing-1

针对“北京一号”多光谱影像存在的因系统影响造成的图像模糊问题与数据量大的特点,对图像的复原方法进行了研究,并提出一种通用的处理流程框架。这种处理流程框架对计算机硬件环境具有良好的灵活性和适应性,为类似大数据量的图像处理提供了一种可供参考的方法。在提高其图像质量方面本文采用了一种快速有效的图像复原算法,即在频率域基于MTF传输调制函数的维纳滤波的复原方法。本文还通过实验对复原算法和流程框架进行了有效性和实用性方面的验证。由于这种解决方案兼顾了复原效果和运算效率,取得了令人满意的结果。

In this paper, a research on the multispectral image restoration of Beijing-1 mierosatellite has been done, mainly concentrated on the blurriness, which is caused by the image system, and the tremendous image data. An universal processing framework has been put forward which has good flexibility and adaptiveness and provides an exemplary method for handling mass image data. The MTF （Modulation Transfer Function）-based Wiener filter, which is comparatively fast and effective, in frequency domain is adopted to improve the image quality. The experiments proved that the restoration method and the framework are practical and effectual as both the result and the computational efficiency were taken into consideration.