河南省冬小麦农田环境参数反演及系统设计

发布时间：2019-02-15 04:34

【摘要】：冬小麦农田环境参数(如地上部分生物量,叶面积指数等)是农业生产的关键信息,是冬小麦生长环境和生长状态的直接反映,与冬小麦的最终产量息息相关。准确及时的获取冬小麦产区的冬小麦种植面积和冬小麦田间参数是进行冬小麦产量预测和农业生产调整的前提。随着3S技术在农业领域的不断发展和应用,利用遥感信息技术提取农作物种植面积,反演农作物农田环境参数成为农业遥感研究的热点。河南省是我国农业大省,其冬小麦生产为我国粮食安全提供了重要保障。利用遥感技术获取河南省冬小麦生产信息具有重要意义。本文以河南省为研究区,以EOS-MODIS和Landsat 8 OLI为数据源,结合野外调查和田间测量数据,综合应用遥感技术和统计分析,实现了河南省冬小麦种植空间信息提取以及田间叶面指数反演和地上生物量反演。并且,在此基础上,基于ENVI/IDL开发了河南省冬小麦农田环境参数反演系统,实现了反演数据的快速处理和反演过程的流程化进行。得到的主要结论如下:(1)采用定量光谱分析方法,基于多时相的MODIS-EVI数据集对河南省主要地物类型的像元EVI值变化情况进行分析表明,提取河南省冬小麦种植空间分布信息的关键时相分别为10月16日、12月9日、4月25日和5月25日,其对应的冬小麦生育期分别为播种期、越冬初期、扬花期和收获期。(2)采用决策树分类算法,基于多时相的MODIS-EVI数据集对河南省冬小麦种植空间分布信息进行提取得到的结果表明,2014年10月到2015年6月这一季河南省冬小麦种种植面积为55484平方千米,与统计面积相比,精度为97.7%,根据实测数据验证得到的用户精度为83.3%。(3)采用统计回归方法,基于实测冬小麦叶面积指数数据和对应的遥感数据进行河南省冬小麦叶面积指数进行反演,结果表明,和冬小麦叶面积指数变化密切的遥感信息特征参量为增强型植被指数EVI和差值植被指数DVI。其中最优的反演模型为LAI=2.844e1.096*EVI,预测模型的决定系数和预测误差分别为0.79和0.319。(4)采用SPSS回归分析功能,结合实测冬小麦地上部分干生物量和对应的遥感植被指数进行反演,结果显示,在冬小麦扬花期和灌浆期,和冬小麦地上部分干生物量变化相关性最强的植被分别为,比值植被指数RVI和归一化植被指数NDVI。两个时期建立的冬小麦地上部分干生物量最优反演模型分别为DM=0.039+0.012×RVI和DM=1.187e1.61NDVI。
[Abstract]:The environmental parameters of winter wheat (such as aboveground biomass, leaf area index, etc.) are the key information of agricultural production and directly reflect the growth environment and growth state of winter wheat, which are closely related to the final yield of winter wheat. Accurate and timely acquisition of winter wheat planting area and winter wheat field parameters is the premise of winter wheat yield prediction and agricultural production adjustment. With the continuous development and application of 3s technology in the field of agriculture, extracting crop planting area by using remote sensing information technology and retrieving agricultural environmental parameters has become a hot spot in agricultural remote sensing research. Henan Province is a big agricultural province in China, and its winter wheat production provides an important guarantee for China's grain security. It is of great significance to use remote sensing technology to obtain winter wheat production information in Henan Province. This paper takes Henan Province as the research area, takes EOS-MODIS and Landsat 8 OLI as data sources, combines field investigation and field survey data, synthetically applies remote sensing technology and statistical analysis. The spatial information extraction of winter wheat planting in Henan Province and the inversion of leaf surface index and aboveground biomass were realized. Furthermore, based on ENVI/IDL, the inversion system of winter wheat farmland environment parameters in Henan Province is developed, which realizes the fast processing of inversion data and the process of inversion. The main conclusions are as follows: (1) quantitative spectral analysis method is used to analyze the changes of pixel EVI values of the main feature types in Henan Province based on the multi-temporal MODIS-EVI data set. The key phases for extracting spatial distribution information of winter wheat planting in Henan Province were October 16, December 9, April 25 and May 25, respectively. (2) decision tree classification algorithm was used to extract spatial distribution information of winter wheat planting in Henan Province based on multi-temporal MODIS-EVI data set. From October 2014 to June 2015, the planting area of winter wheat seeds in Henan Province was 55484 square kilometers, compared with the statistical area, the precision was 97. 7 percent. According to the measured data, the user accuracy is 83.3%. (3) based on the measured winter wheat leaf area index data and the corresponding remote sensing data, the winter wheat leaf area index of Henan Province is retrieved by statistical regression method. The results show that the characteristic parameters of remote sensing information are enhanced vegetation index (EVI) and differential vegetation index (DVI.), which are closely related to the variation of leaf area index of winter wheat. The optimal inversion model is the determination coefficient and prediction error of LAI=2.844e1.096*EVI, prediction model are 0.79 and 0.319.9 respectively. (4) SPSS regression analysis is used. Combined with the measured dry biomass of the aboveground part of winter wheat and the corresponding remote sensing vegetation index, the results showed that the vegetation which had the strongest correlation with the variation of dry biomass of the aboveground part of winter wheat at flowering stage and filling stage of winter wheat were, respectively. Ratio vegetation index RVI and normalized vegetation index NDVI. The optimal inversion models of dry biomass of winter wheat in two periods are DM=0.039 0.012 脳 RVI and DM=1.187e1.61NDVI., respectively.
【学位授予单位】：郑州大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：S512.11;S127

【相似文献】