我国是一个农业大国，受季风和气候的影响，易爆发干旱，给农作物造成了重大的 损失。探索一些高效及时的监测干旱的方法成为了研究干旱的一部分重要工作。遥感具 有高时间分辨率，高光谱分辨率，易获取等特点，由于干旱有“干旱一大片”的特点， 所以不需要非常高的空间分辨率遥感影像。因此使用遥感的方法监测干旱，将大大提高 抗旱水平，为减轻农业灾害服务。干旱产品在应用之前需要进行真实性检验。 遥感的真实性检验就是利用直接检验或间接检验的方法，对拟检验的遥感数据产品 的辐射精度、几何精度和光谱精度进行检验和评价，对遥感反演产品和应用产品的精度 进行检验和评价。本文的重点就是利用目前正在使用的遥感干旱指数，并适当改进，进 行干旱产品的反演，同时使用同时间段的气象干旱指数监测同地区，建立真实性检验同 量纲之间和不同量纲之间的真实性检验流程，并分析三个指数的优缺点以及适用性。 本文以黄淮海平原为研究区域，分别基于遥感指数-植被健康指数（VHI）和改进的 植物水分指数（MNDMI）和气象指数-标准化降水指数（SPI）监测该区域 2000-2012 年的干旱变化情况，并结合研究区域的高程和植被覆盖类型，完成真实性检验的过程。 研究的内容和结论如下： (1) 研究了 VHI 遥感干旱指数的监测原理，并基于 MODIS 的 2000-2012 年遥感影 像数据反演 VHI，最后使用 VHI 评价黄淮海平原的干旱情况； (2) 研究了 NDMI 遥感干旱指数的监测原理，并对 NDMI 进行了距平处理，处理后 的指数称为 MNDMI，MNDMI 考虑了历年的 NDMI 对当年的影响，并基于 MODIS 的 2000-2012 年遥感影像数据反演 MNDMI，最后评价黄淮海平原的干旱情况； (3) 基于 DEM 数据，对黄淮海平原的高程分布进行分析，并将它按照 0-25、 25-50、 50-100、100-500m 的高程进行分类，使用 IDL 将影像 RGB 显示； (4) 基于 MOD12Q1 数据，提取了黄淮海平原的 11 年的植被覆盖情况，将 11 年植 被类型未变的区域仍赋值为原植被类型，将植被类型一年以上改变的区域赋值为 0，最 后使用 IDL 语言将结果以 RGB 的形式表示出来； (5) 研究 SPI 气象干旱指数的监测原理，并使用 IDL 语言实现了该算法，基于 1960-2012 年的降雨数据，计算了 52 年间的 SPI 月值，提取 2001-2012 年间的 SPI 值用 于真实性检验中的不同量纲的检验； (6) 提取 VHI 和 MNDMI 与气象站点相同位置的九像素值， 将 VHI 假设为待检验遥 感产品，用 MNDMI 和 SPI 作为检验产品，实现真实性检验过程，并分析检验结果，确 定三种干旱指数的优缺点及各自适用性。 研究结果表明：VHI 和 MNDMI 对于水分比较敏感，适合实时监测干旱；SPI 适合 分析长时间序列降水跟干旱的关系;MNDMI 监测的干旱规律与 VHI 是一致的，但是根 据本文中描述的分类方法分类的出的监测结果与 VHI 不是很匹配，应该结合更多的实地 数据探索更为适合 MNDMI 的干旱分类标准； 不同量纲的真实性检验结果明显低于同量 纲之间的检验结果，但是对于干旱监测的规律是一致的。

China is a large agricultural country that affected by the monsoon and climate, with the c haracteristics of drought-prone which caused significant losses to crops. An urgent problem i n the drought field that is to find an efficient and timely way for monitoring the drought.Remo te sensing have the features of high temporal resolution, high spectral resolution and easily ob tained and so on. Usually we don’t need high spatial resolution remote sensing images for mo nitoring because drought is a large area case.Therefore, the use of remote sensing methods for monitoring drought, will greatly enhance the level of fight a drought, service at reduce agricultural disaster.Drought product needs to be validated before the application. RS Validation is the use of direct or indirect test, to testing and evaluation the radiation accuracy , geometric accuracy and spectral accuracy of proposed test sensing data, to tested and evaluated the precision of the application and inversion of remote sensing products.The focus of this paper is use of remote sensing drought indices, and appropriate improvements, the inversion of drought products, while the use of meteorological drought index for the same period to monitor the same region, to establish the validation between the same dimension and different dimensions, and to analyze the advantages , disadvantages and applicability of the three indices. In this paper, the North China Plain to the study area, respectively, based on remote sensing index - Vegetation Health Index (VHI) and Modify NDMI and meteorological index -Standardized Precipitation Index (SPI) to monitor changes in the drought years of 2000-2012 in the region and analyze the effects of elevation and vegetation cover types of drought.Finally, complete the process of validation and the applicability of the three indices. The content and conclusions of the study are as follows: （1） Study the monitoring principles of remote drought index VHI, and based on MODIS remote sensing data inversion VHI from 2001 to 2012, Use VHI evaluation of the North China Plain drought conditions, and access to the calendar year drought verifies drought monitoring VHI index. （2） Study the monitoring principles of remote drought index NDMI, and improved NDMI, named modify NDMI, Improved index takes into account the impact on them over the years, and based on MODIS remote sensing data inversion MNDMI, finally evaluation of the North China Plain drought conditions from 2001 to 2012. （3） Based on DEM remote sensing analyzed elevation distribution of the North China Plain, and classified according 0-25,25-50,50-100,100-500 elevation, using IDL to RGB display the image, the final analysis of the effects of different elevations on the drought situation. （4） Based on MOD12Q1 data,Extraction of the 11-year vegetation coverage of the North China Plain, The 11-year vegetation types of the area remains unchanged assign the original vegetation type, the vegetation types more than a year to change assign zero. Finally, analysis of the impact of the drought situation in different vegetation. （5） Study the monitoring principles of drought index SPI, and using IDL achieve algorithm. Based on Rainfall data from 1960 to 2012,SPI values were calculated for 52 years between January. extract SPI values from 2001 to 2012 years to validate different dimension. （ 6 ） Extract Nine pixel VHI and MNDMI values at the same position with the meteorological station. The VHI is assumed to be validated remote sensing products With MNDMI and SPI as a test product, achieve validation process, analyze test results, determine the applicability of advantages and disadvantages of the three drought indices. The results show that: VHI and MNDMI more sensitive to moisture, suitable for real-time monitoring of drought; SPI suitable for analyzing time series of precipitation relationship with drought; MNDMI drought monitoring rule is consistent with the VHI, but according to the classification described in this article, the monitoring results are not a good match with the VHI, Should be combined with more field data to explore more suitable MNDMI drought classification; Different dimensions of validation results were significantly lower than the results between same dimensions, but drought monitoring rule is consistent.