Knowledge Resource Center for Ecological Environment in Arid Area
| 科尔沁沙地荒漠化遥感监测研究 | |
| 其他题名 | STUDY ON DESERTIFICATION MONITORING IN KERQIN USING REMOTE SENSING |
| 廖楚江 | |
| 出版年 | 2007 |
| 学位类型 | 博士 |
| 导师 | 王长耀 |
| 学位授予单位 | 中国科学院遥感与数字地球研究所 |
| 中文摘要 | 荒漠化遥感监测是荒漠化监测研究的重要内容，遥感技术作为地球信息科学的前沿技术，可以在短期内连续获取大范围的地面信息，实现荒漠化信息的快速收集和定量分析，反应迅速，客观性强，是目前最为有效的对地观测技术和信息获取手段。 \n本文将目光聚焦于内蒙古科尔沁沙地，首先基于高精度影像从小尺度上分析该地区典型区域的荒漠化现状及动态变化，而后上升到中分辨率影像进行大尺度的荒漠化监测。\n地质统计学影像纹理考虑到地物空间结构特征的差异，从而在原始光谱波段的基础上增加了空间结构信息，有助于提高遥感荒漠化监测的精度，本文应用该方法开展了一系列荒漠化监测相关的研究。\n论文的主要创新点和成果如下：\n（1）提出了基于地质统计学影像纹理的荒漠化分类方法\n选择非生长期影像监测荒漠化的状态，本文从荒漠化土地的地表结构差异入手，在前人对地质统计学影像纹理的研究基础上，将该方法引入到荒漠化遥感监测中，通过不同荒漠化土地类型的变异函数曲线差异，论证了变异函数纹理在高分辨率影像荒漠化分类中的有益贡献，在具体的纹理计算过程中，本文提出采用四种修正变异函数计算影像纹理，提出采用模型拟合得到变差距离来决定计算窗口可选大小及应用主成分波段作为纹理计算的基础。在对影像纹理的运用上，首先基于三个可选窗口计算影像纹理，而后将它们与原始光谱波段结合起来进行分类，通过比较分类精度选择最适合该类影像纹理计算的窗口大小，最后将典型区域的研究结果推广到科尔沁其它旗县的同类数据。 \n（2）提出了基于变异函数模型参数纹理的荒漠化动态变化监测方法\n 选择生长期影像监测荒漠化土地与其它土地利用类型之间的动态变化，受制于光谱波段和变异函数纹理波段结合起来仍然不能在不同类型耕地和不同级别荒漠化分类中获得理想结果，本文研究提出利用变异函数模型参数纹理提高植被分类的精度，将荒漠化监测的触角延伸到耕地、林地和草场的动态变化监测中，从而能更好地描述荒漠化土地动态变化的来源。\n（3）提出了应用变异函数模型参数的植被群落空间结构特征分析方法\n 鉴于不同植被群落组成对荒漠化类型的指示作用，本文从地质统计学变异函数模型参数中得到启发，将从影像中计算得到的变异函数模型参数跟植被群落空间结构特征对应起来，即计算不同植被群落的实验变异函数并拟合到球状模型，用变差距离和基台值来表达植被群落的空间结构特征。运用这些计算结果，对从典型区域选取的样地进行生态退化和恢复的监测研究并对恢复方法提供一定建议。\n（4）构建了基于MODIS影像的荒漠化监测指标体系\n在前人研究的基础上，提出采用MODIS EVI、Albedo、LST和LSWI四个指标来监测大尺度范围的荒漠化动态变化，四个荒漠化监测指标阈值通过利用高分辨率荒漠化遥感监测研究成果来建立，从而实现了从高分辨率小尺度变换到中分辨率大尺度的荒漠化遥感监测。 |
| 英文摘要 | Remote sensing is an important tool of desertification monitoring, Remote sensing technology is the most effective observing technology and the means of obtaining information at present, we can observe continuous land surface information in a large scale in a very short term, collecting quickly analyzing quantitativly desertification information.\n This paper focuses on Kerqin region in Neimenggu Province, firstly based on high precision image to analysize desertification changes of some typical regions in a small scale, and then raises the research scope, using middle precision image to analysis desertification in a large scale.\nGeostatistical image textures consider the differecnce among differenet land classes, so using this kind of texture will add spacial structure information on base of original spectrum bands, which will be helpful to improve the precision of desertification monitoring. This paper will elaborate on remote sensing desertification monitoring based on geostatistial image texture method. \nThe main innovation points and results production in this paper is as follow：\n(1) Desertification monitoring research using geostatistical image texture\nLauching from the structural difference of desertification land suface, this paper introduced geostatistical image texture into desertification monitoring research, analyzing the difference of variogram among different desertification land classes, proving variogram texture can improve the classification precision of high resolution image, in the calculation of image texture, this paper put forward using four kind of variograms to calculate image texures, and put forward using range to decide the optional window of calculation, and using principle components as the bands of texture calculation. After getting optional windows, we firstly calculate image textures based on three optional windows, and then toghter with original spectral bands to classify typical regions, through comparing the classification precisions, we can find the optimal window of calculating geostatistical image texture that fit to this kind of region, and lastly we can extend the research results to other regions.\n (2) The spacial structure character analysis of vegetation community using variogram model parameter \n This paper selected gowth term images to monitor the dynamic chages between desertification lands with other land use classes. On account of combining spectral bands and texture bands we still could not get perfect classification results toward different plantation classes and grass land classes, this paper brought forward using variogram model parameters to improve the precisions of vegetation classification, broading the scope of desertification monitoring, and monitoring the dynamic changes among plantation, woodland and grass land, consequently, we could use the results to describe the sources of dynamic changes. \n(3) The dynamic changes monitoring research of desertification using variogram model parameter \nBecause different vegetation community constitutes have indications for desertification classes, this paper corresponded variogram model parameters to vegetation community. After calculating vegetation community experimental variogram and fitting spherical model, we used the range and sill to express spacial structural character of vegetation community. By means of these results, this paper researched the degradation and resume of some samples in typical regions. \n(4) The building of desertification monitoring indicators based on MODIS image\n Based on previous researches, this paper put forward using MODIS EVI, Albedo, LST and LSWI to monitor dynamic changes of desertification in large scale, four desertification indications threshold were confirmed through making use of previous results. This paper found a new idea of desertification monitoring from high resolution image and small scale to middle resolution image and large scale. |
| 中文关键词 | 地质统计学纹理 ; 荒漠化监测 ; 空间结构特征 ; 变异函数模型 ; 荒漠化监测指标 |
| 英文关键词 | Geostatistical image texture Desertification monitoring Spacial structural Characteristic Variogram model Desertification monitoring indications |
| 语种 | 中文 |
| 国家 | 中国 |
| 来源学科分类 | 地图学与地理信息系统 |
| 来源机构 | 中国科学院遥感应用研究所 |
| 资源类型 | 学位论文 |
| 条目标识符 | http://119.78.100.177/qdio/handle/2XILL650/286602 |
| 推荐引用方式 GB/T 7714 | 廖楚江. 科尔沁沙地荒漠化遥感监测研究[D]. 中国科学院遥感与数字地球研究所,2007. |
| 条目包含的文件 | 条目无相关文件。 | |||||
| 个性服务 |
| 推荐该条目 |
| 保存到收藏夹 |
| 导出为Endnote文件 |
| 谷歌学术 |
| 谷歌学术中相似的文章 |
| [廖楚江]的文章 |
| 百度学术 |
| 百度学术中相似的文章 |
| [廖楚江]的文章 |
| 必应学术 |
| 必应学术中相似的文章 |
| [廖楚江]的文章 |
| 相关权益政策 |
| 暂无数据 |
| 收藏/分享 |
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。
