【摘要】 小波变换具有数学显微镜特点和频域带通特性 ,可用于对遥感图像进行分析 ,为了探索更好的遥感图像尺度分析方法 ,提出利用二维连续小波变换 (墨西哥帽小波 )结合地物类型分布图来对NOAA/AVHRR影像的 4波段数据进行尺度分析。结果表明 ,在小尺度下连续小波变换系数可显示不同地物类型、相对差异、位置及形状等信息 ,可用作细致分析 ;而在大尺度下该系数则主要表现了由水陆、地貌导致的地域差异 ,可用于概貌观察。另外 ,不同时相数据的大尺度分析对比 ,还体现了空间格局和变化趋势。通常小波变换系数确定的不同地物类型的尺度曲线反映了不同地物的影像信号强度和相互影响 ,而过零点位置的变化则反映了信号的突变更多还原

【Abstract】 Wavelet analysis is chiefly due to the ‘adaptive feature’ and ‘mathematical microtelescope feature’. The 2-D continuous wavelet transform (CWT) is a powerful new tool and has been applied to a number of problems such as astrophysics, aero-magnetic processing, seismic and gravity. It is also used in remote sensing image analysis. This paper focuses on multi-scale analysis at NOAA/AVHRR thermal data (Channels 4 when present). The approach is using an isotropic 2D Mexican hat wavelet(DOG m =2), and studied each component with multi-scale matched on multi-date data against the distribution map of land-cover classification, which is made up to each pixel on NOAA/AVHRR image with TM image supported, to reveal 2D signals in the temporal variation and spatial patterns. In a word, the result shows the information about type of land cover classification and relation, location, and shape in its in small scale as micro-scale analysis and emphasize terraqueous variance by physiognomy, the strengths and features of its trend and structure in large scale as macro-scale observer. The variance of coefficient of different land cover types and the zero-crossing variance of coefficient with scale in 2D CWT discovers the power of signal and the correlation.更多还原