图像信息容量与区域下垫面的空间结构特征研究

发布时间：2018-05-01 02:27

本文选题：遥感图像 + 信息容量　；参考：《西北大学》2013年硕士论文

【摘要】：图像信息容量源于计算机图形学,它是图像像素值之间差异程度的反映,是图像有效信息承载力的数量单位。遥感图像信息容量模型构建是基于图像多维直方图理论,其数值能够简易、直观的反映地表地物的复杂程度,其数值大小主要受到地貌类型、地表覆盖、地物结构特征等因素的影响。本文在现有图像信息容量研究的基础上,以陕西省关中平原地区为例,以Landsat5TM多光谱数据为主要数据源,采用空间分析与数理统计相结合的方法,重点分析了城镇区域下垫面组分以及遥感图像信息容量的空间结构特征。本文研究主要取得以下几项结论： (1)在传统遥感图像信息容量计算的基础上,提出了基于滑动窗口的信息容量计算模型,并将其应用于实验中,分析了计算结果——窗口信息容量图像的空间分布特征。 (2)文章以尺度变换为分析手段,借鉴景观格局研究方法,通过对多种结构指数在各粒度层次以及幅度范围上响应分析,得到文章所选区域下垫面结构研究的最适粒度和幅度,约为40m和18km。 (3)论文采用混合像元光谱线性分解技术提取了该区域各下垫面组分,并分别以定性描述与定量模拟的方式探讨了图像信息容量以及下垫面组分的空间分布结构特征。结果表明,城镇区域下垫面组分中植被覆盖以及不透水面是影响遥感图像信息容量数值分布的主要因素,且几者在空间分布上存在着较强的敏感性,即植被覆盖以及不透水面的高值区域均对应于信息容量的高值区域,呈现较高的正相关性关系,其相关性系数分别为0.7517和0.6849。
[Abstract]:Image information capacity is derived from computer graphics. It reflects the degree of difference between pixel values of images and is the quantitative unit of the effective information carrying capacity of images. The information capacity model of remote sensing image is based on the theory of multi-dimensional histogram. Its numerical value can easily and intuitively reflect the complexity of surface objects, and its numerical value is mainly affected by the geomorphological type and surface coverage. Features of the structure of features and other factors. Based on the existing research of image information capacity, taking Guanzhong Plain area of Shaanxi Province as an example, taking Landsat5TM multispectral data as the main data source, this paper adopts the method of combining spatial analysis with mathematical statistics. The spatial structure characteristics of the underlying surface components and the information capacity of remote sensing images in urban areas are analyzed. The main conclusions of this study are as follows: 1) on the basis of traditional remote sensing image information capacity calculation, a sliding window based information capacity calculation model is proposed, and it is applied to the experiment. The spatial distribution characteristics of the calculated result-window information capacity image are analyzed. (2) by using scale transformation as an analysis method and using the research method of landscape pattern for reference, by analyzing the response of various structural indices at various granularity levels and amplitude ranges, the optimum granularity and amplitude of the study of underlying surface structure in the selected area are obtained. It is about 40m and 18km. In this paper, the components of the underlying surface in the region are extracted by using the mixed pixel spectral linear decomposition technique, and the information capacity of the image and the spatial distribution structure of the underlying surface components are discussed by qualitative description and quantitative simulation, respectively. The results show that vegetation cover and impermeable surface are the main factors affecting the numerical distribution of the information capacity of remote sensing images, and some of them have strong sensitivity in spatial distribution. That is, the high value region of vegetation cover and impermeable water surface corresponds to the high value region of information capacity, and the correlation coefficient is 0.7517 and 0.6849, respectively.
【学位授予单位】：西北大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P208

【参考文献】