针对风云二号(FY-2)中波红外波段地球目标图像空间分辨率低的问题, 提出了从月球图像中提取刃边目标进行调制函数(MTF)计算, 根据图像质量评价参数进行图像复原提高图像空间分辨率的方法。首先根据FY-2定位模型设计了月球可视位置算法, 搜索到月球图像; 通过一阶微分、极值匹配等处理, 基于傅里叶变换得到FY-2G中波红外波段在轨MTF评价曲线。然后, 根据MTF计算结果, 采用维纳滤波器对中波红外波段图像进行复原; 以像素平均梯度、功率谱分量和以及图像信息熵作为评价参数, 通过3种评价参数的变化确定合理的滤波参数。最后, 得到经过MTF复原的高分辨率图像。实验结果显示, 应用提出的方法, 不仅MTF计算正确, 图像复原有效, 空间分辨率得到提高, 还保证了图像复原前后能量的一致性。

For improving the spatial resolution of FY-2G middle infrared band, this paper proposes a method to extract the edge from a moon image and to calculate the Modulation Transfer Function(MTF). It implements image restoration based on the image quality evaluation parameters and improves the image space resolution. Firstly, a moon visual position algorithm was designed based on the image navigation of FY-2 to search the moon image in the FY-2G full disk image. By using one-order differential and extreme value matching, the MTF evaluation value of FY-2G middle infrared band was calculate out by Fourier transform. Using the calculated MTF, the full disk image was restored by a Wiener filter. By using the pixel average gradient, power spectral components and image information entropy as evaluation functions, the reasonable filter parameters were determined. Finally, the high resolution image after MTF restoration was obtained. The experiments indicate that the MTF calculation is correct, image restoration is effective, and the spatial resolution has been improved greatly. Moreover, it ensures the consistency of image energy before and after restorations.