超分辨率重构技术用于红外扫描系统的关键

摘要

超分辨率重构(SRR)技术是提高遥感图像空间分辨率最有前景的方法,红外遥感成像的空间分辨率相对可见光要低得多,研究采用SRR技术提高红外扫描图像空间分辨率有重要意义.从有效应用SRR技术的条件出发,分析它应用于红外扫描成像系统需要解决的关键问题,对待重构图像错位量估计可通过采用固定错位量的多排传感器,并精确计算其曝光时间来解决；对待重构图像中噪声的去除,提出了基于序列逐像素比较滤波法.通过红外扫描成像实验,提出的方法可以有效解决待重构图像错位量估计不准与噪声对重构的影响.%Super-Resolution Reconstruction (SRR) is the most prosperous method for improving the spatial resolution of remote sensing images. The spatial resolution of remote sensing images in infrared band is much lower than that in visible band. Based on the application conditions of SRR, the critical problems applying SRR in infrared scanning imaging system were analyzed. The displacements among the reconstructed low-resolution images were estimated by using the multiple rows of sensors with fixed displacement and calculating precisely the scanning integral time. The noises of low-resolution images were removed by filtering pixels in the relative location on image sequences one by one. The result of infrared scanning imaging experiment proved that the proposed method can effectively deal with the displacement estimation and noise reduction in reconstructed low-resolution images.