基于风云气象卫星数据的内蒙古草地干旱监测

发布时间：2018-04-17 12:18

本文选题：草地 + 分类　；参考：《电子科技大学》2017年硕士论文

【摘要】：干旱一直是对中国影响最主要的气象灾害之一,它对人类生存、生活的环境和经济、社会的发展,都有着严重的负面影响。内蒙古草原面积辽阔,草地干旱是内蒙古主要的自然灾害之一,因此开展内蒙古草地干旱灾害监测,建立有效的草原干旱监测模型,具有一定的理论意义和现实价值。本文基于风云三号气象卫星(FY-3)卫星的中分辨率光谱成像仪(Medium Resolution Spectral Imager,MERSI)数据和可见光红外辐射计(Visible and Infra-Red Radiometer,VIRR)数据,开展内蒙古地区草地干旱监测研究。主要研究工作如下:(1)内蒙古地区草地分布提取。经过对比三种分类方法的精度及适应性,选取随机森林分类方法提取内蒙古地区草地分布。首先筛选出图像质量好的MERSI数据,分别采用随机森林、最大似然法、决策树法进行分类,利用美国陆地卫星8号星(Landsat8)影像数据对结果进行精度验证;分别对三种分类方法作精度分析对比后,选取随机森林法对草地进行分类提取,做出本论文研究所需的草地分布图。(2)草地干旱监测指数研究。利用MERSI数据产品生成归一化植被指数(Normalized Difference Vegetation Index,NDVI)和地表温度(Land Surface Temperatrue,LST)旬产品后,分别对基于温度植被指数(Temperature Vegetation Dryness Index,TVDI)、植被供水指数、热惯量指数模型进行反演以及计算,生成旬际的干旱指数产品结果图,并结合草地分布图,生成草地干旱图;并对比三种指数的反演结果精度。实验结果表明基于TVDI的草地干旱监测精度最高。(3)基于TVDI的草地干旱分级。结合内蒙古草原旱情情况,建立了适合内蒙古草原的干旱指数分级标准,生成干旱等级示意图;并结合气象资料对内蒙古草原的空间、旬月年际的变化进行分析。以上工作表明利用FY-3A/MERSI卫星资料监测内蒙古草原干旱具有可行性。
[Abstract]:Drought has always been one of the most important meteorological disasters affecting China, which has a serious negative impact on human survival, living environment, economy and social development.Inner Mongolia has a vast grassland area, grassland drought is one of the main natural disasters in Inner Mongolia, so it is of theoretical significance and practical value to carry out the monitoring of grassland drought disasters in Inner Mongolia and to establish an effective grassland drought monitoring model.Based on the medium Resolution Spectral imager MERSI data of FY-3 and visible and Infra-Red Radiometeror VIRR data, the grassland drought monitoring in Inner Mongolia is carried out.The main research work is as follows: 1) extraction of grassland distribution in Inner Mongolia.By comparing the accuracy and adaptability of the three classification methods, the random forest classification method was selected to extract the grassland distribution in Inner Mongolia.Firstly, the MERSI data with good image quality are selected and classified by random forest, maximum likelihood method and decision tree method, and the accuracy of the results is verified by the image data of American Landsat 8 satellite.After analyzing and comparing the accuracy of the three classification methods, the random forest method was selected to classify and extract the grassland, and the grassland distribution map. 2) the grassland drought monitoring index was studied in this paper.The normalized Difference Vegetation index (NDVI) and Land Surface temperature index (LST) were generated by using MERSI data products. The models of temperature vegetation index, vegetation water supply index and thermal inertia index were retrieved and calculated respectively based on the temperature vegetation index (TVI), the vegetation Vegetation Dryness index (VDI), the vegetation water supply index (VSI) and the caloric inertia index (TCI).The drought index product result map of ten days is generated, and the grassland drought map is generated by combining the grassland distribution map, and the accuracy of the inversion results of the three indices is compared.The results show that the precision of grassland drought monitoring based on TVDI is the highest.According to the drought situation of Inner Mongolia grassland, the drought index classification standard suitable for Inner Mongolia steppe was established, and the map of drought grade was generated, and the spatial and interannual variation of Inner Mongolia grassland was analyzed with meteorological data.The above work shows that it is feasible to use FY-3A/MERSI satellite data to monitor grassland drought in Inner Mongolia.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：S812.1

【参考文献】