节点文献
遥感土地利用动态监测变化信息自动提取与实证研究
Remote Sensing Monitoring Land Use Change Information Automatically Extracted and Empirical Study
【作者】 王瑞国;
【导师】 侯淑涛;
【作者基本信息】 东北农业大学 , 农业遥感与土地利用, 2007, 硕士
【摘要】 遥感土地利用动态监测在我国已有近20年的发展历程,其目的:一是利用遥感影像监测土地利用的动态变化:二是利用监测的变化信息分析影响土地利用变化的因素与变化过程,预测其变化、演化趋势,以指导政府对土地利用管理与决策。遥感土地利用动态监测反映的是区域土地利用动态变化的过程,监测数据呈现出明显的时态特征,随着我国经济建设的迅猛发展,土地利用变化显著,遥感土地利用动态监测周期越来越短,这就需要信息提取快速、高效,并在GIS系统中统一管理和查询分析。传统土地利用动态监测变化信息的提取主要是依据目视解泽与计算机识别交互进行,这种方法耗时耗力,其准确度依赖参与解译分析人的知识与经验,操作程序不具有重复性。遥感变化信息自动提取方法已有多人研究,但大多研究的重点体现在不同方法所适宜的应用领域,很少有人关心各种方法实践应用的难易程度和和变化信息的分类解译。本文根据遥感土地利用动态监测的特点,在总结国内外研究遥感信息提取方法的基础上,系统地阐述了遥感影像融合理论与方法,利用ERDAS、ArcGIS、MAPGIS等软件完成影像增强处理、变化信息提取等研究工作。对常用的图像融合:IHS变换融合、BROVEY变换、小波变换和IHS变换相结合融合、PCA变换融合以及PCA与HIS变换相结合融合等方法进行了试验对比分析与效果评价,建立了信息熵、平均梯度、方差等7个影像融合质量评价指标并对各融合影像效果加以评价,得出IHS变换融合方法效果最好,PCA+IHS变换融合方法次之的结论。在此基础上,进一步研究了变化信息提取方法,并对差值法、比值法、主成分变换法、特征变异法等信息提取方法进行了试验,通过对试验结果的对比分析认为:采用差值法进行变化信息提取的同时,兼顾特征变异法对影像变化信息进行修正,能够提高信息提取的效率,大大减少了变化信息提取的错提、遗漏情况。在上述理论研究的基础上,并以某城市作为研究试验区,利用两时相的SPOT5影像为数据源进行实证研究,建立了遥感土地利用动态监测变化信息自动提取的实践应用框架体系,实现土地利用动态监测变化信息自动提取,达到了遥感土地利用动态监测变化信息提取实践应用的简单化、实用化,减轻作业强度,提高了工作时效,为快速、准确开展土地利用动态遥感监测提供了可行的技术方案,具有一定的实践应用价值。
【Abstract】 Application of remote sensing technology for land use monitoring in our country has beennearly 20 years. There are two purposes for dynamic monitoring: First, using remote sensingimages to monitor land-use changes; another is using monitoring data to analyze the factorsaffecting the process of land-use changes and trends, to guide government’s management andpolicy. In addition, dynamic monitoring reflects the dynamic changes process of region land use,and the temporal characteristic of the monitoring data is obvious. With the development of ourcountry, the changes of land use are greatly, the periods of dynamic monitor get shorter and shorter.Therefore, we need extract information, which is unified managed and queried in GIS, in a rapidand effective way.For traditional methods of land use dynamic monitoring, the extraction of changedinformation mainly depends on visual interpretation and computer synchronization. This methodtakes a lot of time and energy, moreover it relies on the knoeledge and experience of the personwho participate in the visual interpretation. Generally the operation does not need duplication.Although many people have studied the methods of extracting remote sensing changed informationautomatically, most methods focus on the application which is suitable for different fields, and fewconcerned with the degree of realization and the convenience of operation as well as the classifyinterpretion of changed information.My paper is on the base of the development domestically and abroad, viewing of the land usedynamic monitor characteristic.At the same time, my paper has a systerms elaboration about thetheory and method of remote sensing images fusion. I use ERDAS、ArcGIS、MAPGIS to completethe images enhancement and changed information extraction.For the commonly images fusionmethods such as IHS transformation fusion method、BROVEY transformation、the conmbinedfusion of wavelet and IHS transformation、PCA change fusion and the combined fusion betweenPCA and IHS, I carry on experimental comparative analysis and effects appraise, and establishseven images fusion quality evaluation indicators about the information entropy, the averagegradient、variance and so on,as well as evaluate the effection of fusion images. At last by the fusionimage performance index, I come to the conclusion that the fusion method of IHS transformation isbest for dealing with land use dynamic monitor data fusion. The next is PCA+IHS transformationfusion method.On the base of fusion method research, I has further studied the method of changed information extraction, and did comparative experiment on D-value method、radio method, PCA,spectral characteristic variation. Through the experimental comparative analysis, I come to aconclusion that we can improve the information extraction efficiently, reduce the chance of wrongand missing information when use D-value method to extract changed information as well as usespectral characteristic variation to modify image changed information. On the base of the theorystudied above,I selected a city as experimental area, done empirical study by SPOT5 Imagesdata, And I establish the practice application frame system in the foundation of remote sensing landuse dynamic monitor, also realize remote sensing land use dynamic monitoring changedinformation extraction automatically.At last, I acheive the goal that the practice application issimplified, the application is practical, reducing the work intensity, enhancing the workingefficiency. I offer a technical project to make the land use dynamic remote sensing monitor fast andaccurate more and more, so my paper does have practical application value.
【Key words】 Land use dynamic monitoring; changing information; image fusion; Information extraction;
- 【网络出版投稿人】 东北农业大学 【网络出版年期】2008年 02期
- 【分类号】S159.2
- 【被引频次】16
- 【下载频次】1218