基于小波分析的TM遥感图像超分辨率重建  

Super Resolution Reconstruction of TM Remote Sensing Image based on Wavelet Analysis

作　　者：王金亮[1] 曾浩 王艳英[1] 刘广杰[1] 

机构地区：[1]云南师范大学旅游与地理科学学院,云南昆明650500 [2]上海航天电子技术研究所,上海201109

出　　处：《遥感技术与应用》2016年第3期476-480,共5页Remote Sensing Technology and Application

基　　金：国家自然科学基金项目(41271230;40861009);云南省中青年学术技术带头人培养项目(2008PY056)资助

摘　　要：基于图像算法的超分辨率重建技术可以提高光学遥感图像的空间分辨率,能够更加有效地利用现有数据并降低成本。以滇西北香格里拉市小中甸坝为实验区,以2009年的TM影像为数据源,开展遥感图像超分辨率重建实验研究。首先分析其中造成图像退化的各项因素并经过双线性插值、维纳逆滤波、卷积等处理;然后通过小波分解得到描述各个方向上不同尺度的高低频信息的小波系数,并多次试验推导出满足预期条件的加权因子。再将多时段的低分辨率图像小波系数以小波重构的方式重建。通过实验可以看出,重建图像能提供更多的细节信息,图像质量有了明显提高。Super resolution reconstruction algorithm can improve the spatial resolution of optical remote sensing images,which can effectively utilize existing data and reduce costs.Take Xiaozhongdian basin of Shangri-La City in northwest Yunnan as experimental areas and Landsat TM image（2009）as data source,super resolution reconstruction algorithm has been studied in the research.First,we analyzed the various factors of image degradation and processed the image by using bilinear interpolation,Wiener inverse filter,convolution and other treatment.Then got wavelet coefficients on different scales of high/low frequency information by using wavelet decomposition,and derived weighted factor that can meet the expected conditions through many tests.And wavelet coefficient of multi-period low-resolution images has been rebuilt by means of wavelet reconstruction.Experimental results show that the reconstructed image can provide much more detail information and the image quality has been improved significantly than unreconstructed image.

关 键 词：超分辨率重建 TM图像 小波分析 香格里拉 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]