EGI OpenIR
玛纳斯河流域和渭库三角洲植被覆盖的提取技术应用及动态分析
Alternative TitleExtraction technology application and dynamic analysis of vegetation cover in Manas river basin and Ugan-Kuqa river delta oasis
张小溪
Subtype硕士
Thesis Advisor杨辽
2019-06-30
Degree Grantor中国科学院大学
Place of Conferral北京
Degree Discipline工程硕士
Keyword干旱半干旱区 植被覆盖度 玛纳斯河流域绿洲 渭-库三角洲绿洲 时空变化 Arid and Semi-arid Areas Vegetation Coverage Manas River Basin Oasis Ugan-Kuqa River Delta Oasis Time and Space Changes
Abstract植被是地球上不可或缺的部分,也是关乎人类生存的重要物质, 在生态系统中具有十分重要的地位, 它与周围生态环境密切相关, 是反应区域生态环境和全球气候变化的最好的标志之一(孙红雨 等, 1998),对防风固沙、 抑制土地荒漠化、减少水土流失等方面具有显著的作用(詹小国, 2001; Wang, 2002; 叶朝霞 等,2007), 特别是干旱半干旱地区,植被在其生态系统中占据着不可忽视的重要地位。 在植被比较稀疏或者没有植被的区域, 水土流失严重, 导致区域生态系统极其脆弱和不稳定, 荒漠化不断加剧, 盐碱土地的存在使得农作物无法生长, 导致粮食减产,严重危害着粮食安全。对于新疆来说, 由于植被分布十分稀疏,在一定程度上促使了荒漠化的发展,对人类的生产生活和区域生态环境产生了不良影响和威胁。在土地退化比较严重的干旱半干旱地区,植被覆盖是反映区域荒漠化、区域土地退化等的重要指标之一。多年来, 经过专家学者的大量研究研究发现,荒漠地区的植被覆盖不仅能够作为区域当前和近期土地退化的衡量标准,而且也是预测土地退化的比较敏感的指标。因此,对植被覆盖的提取和动态分析,是我国干旱半干旱地区生态环境研究的重要内容之一。本文以 Landsat 系列卫星影像作为数据源,选择新疆北部的玛纳斯河流域绿洲和天山南部的渭库三角洲作为研究区,选用像元二分模型、改进的三波段最大梯度差法、基于光谱归一化的混合像元分解模型对研究区的植被覆盖进行提取,通过对这三个方法的对比评估,选取一种最适合研究区的植被覆盖提取的方法,然后对研究区 2000 到 2018 年的植被覆盖度的时空变化进行分析。主要研究结果与结论如下:(1) 使用了像元二分模型、改进的三波段最大梯度差法、 基于光谱归一化的协同稀疏回归算法对玛纳斯河流域绿洲和渭库三角洲绿洲进行植被覆盖的提取, 通过对三个模型的提取精度的对比评估, 结论发现在两个研究区, 基于光谱归一化的协同稀疏回归模型算法反演的植被覆盖的精度高于另外两个模型。(2) 利用基于光谱归一化的协同稀疏回归算法对玛纳斯河流域绿洲2000-2018 年植被覆盖度的时空变化的分析,得到的结论为: 在时间上, 绿洲的植被覆盖度总体呈现出扩张的变化特征, 在 2010 年的时候出现过比较大面积的植被覆盖的减少, 但随后又开始增加。 在空间上绿洲靠近天山山脉的区域, 植被基本上常年都呈现出连片的生长状况, 随着时间的推移, 绿洲的东北部区域的植被开始向外扩张, 植被覆盖情况有明显的改善。(3) 利用基于光谱归一化的协同稀疏回归算法对渭库三角洲 2000-2018 年植被覆盖度的时空变化的分析,得到的结论为: 在时间上, 渭库三角洲的植被覆盖度上呈现出波动增加的变化, 总体来说和玛纳斯河流域的绿洲的情况一样, 呈现出扩张的变化特征。 在空间上, 越靠近渭干河库车河,植被生长越好, 绿洲的植被覆盖度越高, 随着时间的推移,绿洲由内部向外部呈现出辐射生长的状况。
Other AbstractVegetation is an indispensable part of the earth and plays an important role in theecosystem. Especially for desert ecosystems, vegetation used to indicating changes inits ecological environment, regulating climate, protecting wind and stabilizing sand.The distribution of vegetation on Earth has obvious temporal and spatial distributioncharacteristics. In arid and semi-arid regions, extracting and analyzing thecharacteristics of temporal and spatial changes in vegetation coverage has importantpractical significance for preventing, controlling desertification and revealing changesin ecological environment. Regarding the research of vegetation coverage extraction,experts and scholars have carried out a large number of research and experimentsmany years ago, and have obtained a lot of valuable experience and technical methods.The earliest extraction of vegetation coverage was done by manual measurement. Thistraditional method is time-consuming and laborious, and often results in low precision.Therefore, experts and scholars in related fields have been working harder to researchon the more efficient and more accurate method of vegetation coverage extraction.Then the development of satellite remote sensing brought a new world to theextraction of vegetation coverage. At the same time, with the continuous improvementof time resolution, spatial resolution and spectral resolution of remote sensing data,and the continuous enrichment of satellite data sources, the extraction of vegetationcoverage brings rich data sources and high-quality remote sensing data. The relevantresearch and applications in this field have made great progress.In arid and semi-arid regions, due to sparse vegetation distribution, thesensitivity of satellite remote sensing to detect ground targets is reduced. Comparedwith other areas, the same extraction model in arid and semi-arid areas often getworse results. So the seemingly general approach to vegetation coverage extraction islikely to lose its universality in arid and semi-arid regions. This is a difficult problemin the development of vegetation cover extraction technology. Finding an extraction model suitable for arid and semi-arid areas where have sparse vegetation is crucial toimprove precision, and is one of the research directions in current.This paper takes these question as the point of setting up the question. Firstly,discusses the significance of this research. Then summarizes the research status ofrelated fields in China and at abroad, and understands the problems existing in thecommonly used vegetation coverage extraction model in arid and semi-arid regions.Through the comparative experimental study of the model, it is expected to obtain anextraction model suitable for the extraction of vegetation coverage in arid andsemi-arid regions.In this paper, the medium-resolution multi-spectral imagery is used as the datasource, and the manas River Basin oasis in the northern part of Xinjiang and theUgan-Kuqa river delta oasis in the southern Tianshan Mountains are selected as theresearch area. After a large number of models have been tested, the final selection isused for extraction research. The vegetation coverage is the most conventional,high-precision and computationally efficient model: the pixel dichotomy model, theimproved three-band maximum gradient difference method, and the synergistic sparseregression algorithm based on the spectral normalization framework. In the case ofthe distribution of vegetation in the north and the south is significantly different, bycomparing and analyzing the extraction results of the three methods in the two studyareas, an optimal model is finally determined to be used to extract the vegetationcoverage of the study area, and then based on the vegetation of the study area from2000 to 2018, the temporal and spatial variations of coverage are analyzed. The mainfindings and conclusions are as follows:(1) Using the pixel dichotomy model, the improved three-band maximumgradient difference method, and the synergistic sparse regression algorithm based onspectral normalization to extract the vegetation coverage of the manas river basinoasis and the Ugan-Kuqa river delta oasis. Comparing the extraction accuracy of thethree models, it is found that the accuracy of vegetation coverage inversion based onthe spectral normalization collaborative sparse regression model algorithm is higherthan the other two models in the two study areas.(2) Using the synergistic sparse regression algorithm based on spectralnormalization to analyze the temporal and spatial variation of vegetation coverage inthe manas River Basin oasis from 2000 to 2018, the conclusion is: in time, thevegetation coverage of the oasis is generally presented. Out of the changingcharacteristics of the expansion, there was a relatively large reduction in vegetationcover in 2010, but then began to increase. In the area where the oasis is close to theTianshan Mountains in the space, the vegetation basically shows contiguous growththroughout the year. Over time, the vegetation in the northeastern part of the oasisbegins to expand outward, and the vegetation coverage is obviously improved.(3) Using the synergistic sparse regression algorithm based on spectralnormalization to analyze the temporal and spatial variation of vegetation coverage inthe Ugan-Kuqa river delta oasis 2000 to 2018, the conclusion is: in time, thevegetation coverage of the Ugan-Kuqa river delta oasis Delta is presented. Theincrease in volatility, in general, is similar to the case of the oasis in the manas RiverBasin, showing the characteristics of expansion. In terms of space, the closer to theKuqa River in the Ugan River, the better the vegetation growth, and the higher thevegetation coverage of the oasis. As time goes by, the oasis shows radiation growthfrom the inside to the outside.
Subject Area测绘工程
Language中文
Document Type学位论文
Identifierhttp://ir.xjlas.org/handle/365004/15374
Collection中国科学院新疆生态与地理研究所
研究系统
Affiliation中国科学院新疆生态与地理研究所
First Author Affilication中国科学院新疆生态与地理研究所
Recommended Citation
GB/T 7714
张小溪. 玛纳斯河流域和渭库三角洲植被覆盖的提取技术应用及动态分析[D]. 北京. 中国科学院大学,2019.
Files in This Item:
There are no files associated with this item.
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[张小溪]'s Articles
Baidu academic
Similar articles in Baidu academic
[张小溪]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[张小溪]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.