吉林一号卫星视频影像超分辨率重建

发布时间：2018-04-03 19:53

本文选题：卫星视频　切入点：超分辨率重建　出处：《国土资源遥感》2017年04期

【摘要】：针对吉林一号(Jilin-1)卫星视频影像的超分辨率重建问题,在分析卫星视频与普通视频及遥感影像不同特性的基础上,研究了在成像场景内含有运动物体情况下,Vandewalle方法和金字塔LK(Lucas-Kanade)光流法运动估计方法的适用性。同时,针对目前中值平移并添加(median shift and add,MSA)方法在视频卫星影像重建中因帧间缺少互补信息而引起的边缘信息不清晰问题,提出了新的中值平移并添加(new median shift and add,NMSA)方法。首先,根据重建的倍数建立重建前、后的空间分辨率格网,并将2个格网统一到1个空间内;然后,以待确定的重建后影像格网像元为中心,确定参与估计的低空间分辨率像元值,利用容许误差确定重建后的像元值。通过利用吉林一号(Jilin-1)卫星数据进行实验,验证了NMSA方法的有效性。
[Abstract]:Aiming at the problem of super-resolution reconstruction of Jilin-1 satellite video image of Jilin No. 1, this paper analyzes the different characteristics of satellite video, ordinary video and remote sensing image.The applicability of the Vandewalle method and the LKK Lucas-Kanade-based optical flow method for motion estimation in imaging scenes with moving objects are studied.At the same time, aiming at the problem of indistinct edge information caused by the lack of complementary information between frames in video satellite image reconstruction, a new method of median translation and adding new median shift and add NMSAs is proposed.First of all, according to the multiple of reconstruction, the spatial resolution grid before and after reconstruction is established, and the two grids are unified into one space, and then, the pixel of the reconstructed image grid is taken as the center.The pixel value of low spatial resolution is determined, and the reconstructed pixel value is determined by the allowable error.The validity of NMSA method is verified by using Jilin-1 satellite data.
【作者单位】：辽宁工程技术大学测绘与地理科学学院;黑龙江地理信息工程院;
【基金】：国家自然科学基金项目“MRF模型的车载全景视觉位姿估计最优化方法研究”(编号:41501504) 2016年辽宁省教育厅一般项目“复杂运动场景下卫星视频的超分辨率重建方法研究”(编号:LJYL011)共同资助
【分类号】：TP751

【相似文献】