基于Landsat8 OLI数据彰武地区旱情监测模型研究

发布时间：2018-11-10 17:49

【摘要】：本研究的研究范围是辽宁省阜新市彰武县,运用ENVI软件对研究区2016年6月和4月的Landsat 8卫星OLI遥感影像图进行预处理。同时,获取研究区相应月份的实测土壤含水率数据。本研究选取了温度植被指数法和垂直干旱指数法作为旱情监测的指标,将各指标分别与土壤含水率进行拟合,运用SPSS软件进行显著性分析,最终得出适用于彰武地区的旱情监测模型。本研究的主要内容和结论如下:1.对基于温度植被指数的模型进行构建和分析。通过对比分析温度植被指数与土壤含水率相关性和显著性,结果表明:温度植被干旱指数与不同深度的土壤含水率拟合效果较差,其中与30cm 土壤含水率的拟合效果相对较好,其复相关系数为0.3837;利用SPSS软件对二者的显著性进行分析,分析结果表明,土壤深度为10cm和20cm的土壤含水率与TVDI在0.05水平上显著相关,深度为30cm的土壤含水率与TVDI在0.01水平上显著相关。2.对基于垂直干旱指数的模型进行构建和分析。在基于垂直干旱指数法旱情监测模型构建中,由PDI反演数据与土壤含水率的拟合结果可知,深度20cm的土壤含水率拟合效果最好,复相关系数为0.5555,10cm与30cm深度的土壤含水率拟合效果相对较差。显著性分析结果表明,土壤深度为1Ocm和20cm的土壤含水率与垂直干旱指数在0.01水平上显著相关,深度为30cm的土壤含水率与垂直干旱指数在0.05水平上显著相关。3.对两个模型进行对比分析,选择最优模型作为彰武地区的旱情监测模型。由两个模型的相关性和显著性对比分析结果可知,垂直干旱指数与土壤含水率的拟合结果要好于土壤含水率与TVDI的拟合结果。因此,基于垂直植被指数法的旱情监测模型更适用于彰武地区的旱情监测。4.对构建的模型进行验证分析。通过对2016年4月份卫星影像图进行反演,得到土壤含水率反演值,通过对不同深度实测土壤含水率与反演出来土壤含水率的对比分析,结果表明,深度为1Ocm和20cm的土壤含水率的相关性分别高出深度为30cm 土壤含水率相关性0.1041和0.1064。其中,深度为30cm的土壤含水率相关性最差,其相关系数为0.4945;深度为10cm的反演土壤含水率与实测土壤含水率的复相关系数为0.5986;深度为20cm的反演与实测土壤含水率的复相关系数为0.6009。通过精度和相对误差分析,10cm、20cm反演效果最理想,10cm反演精度平均值为91.25%,20cm反演精度平均值为88.53%,分别高出30cm反演精度平均值6.96%和4.24%。在相对误差分析中10cm、20cm的相对误差平均值分别低于30cm相对误差8.9%和5.34%,30cm的相对误差平均值偏大为18.49%。综上所述,在旱情监测中,运用遥感手段和垂直干旱指数模型进行监测时,深度小于20cm的干旱监测效果比较理想。
[Abstract]:The research scope of this study is Zhangwu County, Fuxin City, Liaoning Province. The Landsat 8 satellite OLI remote sensing images in June and April 2016 are preprocessed by ENVI software. At the same time, the measured data of soil moisture content in corresponding months in the study area were obtained. In this study, the temperature vegetation index method and vertical drought index method were selected as indicators of drought monitoring. The indexes were fitted with soil moisture content, and the significance was analyzed by SPSS software. Finally, a drought monitoring model suitable for Zhangwu area is obtained. The main contents and conclusions of this study are as follows: 1. The model based on temperature vegetation index is constructed and analyzed. The correlation and significance between temperature vegetation index and soil moisture content were compared and analyzed. The results showed that the fitting effect of temperature vegetation drought index and soil moisture content at different depths was poor, and the fitting effect with 30cm soil moisture content was relatively good. The multiple correlation coefficient was 0.3837; The significance of the two was analyzed by using SPSS software. The results showed that the soil moisture content with the depth of 10cm and 20cm was significantly correlated with TVDI at the level of 0. 05. Soil moisture content with depth of 30cm was significantly correlated with TVDI at 0.01 level. The model based on vertical drought index is constructed and analyzed. In the construction of drought monitoring model based on vertical drought index method, the fitting results of PDI inversion data and soil moisture content show that the soil moisture content fitting effect of depth 20cm is the best. The fitting effect of soil moisture content with the depth of 30cm was relatively poor when the complex correlation coefficient was 0. 5555 ~ 10 cm and the depth of 30cm was 10 cm. The results of significant analysis showed that soil moisture content with 1Ocm and 20cm depth was significantly correlated with vertical drought index at 0.01 level, and soil moisture content with depth 30cm was significantly correlated with vertical drought index at 0. 05 level. The two models were compared and the optimal model was chosen as the drought monitoring model in Zhangwu area. The correlation and significance of the two models showed that the fitting results of vertical drought index and soil moisture content were better than that of soil moisture content and TVDI. Therefore, the model based on vertical vegetation index method is more suitable for drought monitoring in Zhangwu area. 4. The model is validated and analyzed. Through the inversion of satellite image in April 2016, the inversion value of soil moisture content is obtained. Through the comparison and analysis of soil moisture content measured at different depths and the inversion soil moisture content, the results show that, The correlation of soil moisture with depth of 1Ocm and 20cm was higher than that of 30cm with depth of 0.1041 and 0.1064, respectively. Among them, the correlation coefficient of soil moisture content with depth of 30cm is the worst, the correlation coefficient is 0.4945, the complex correlation coefficient between soil moisture content and measured soil moisture content is 0.5986 when the depth is 10cm, and the correlation coefficient is 0.4945 when the depth is 30cm. The complex correlation coefficient between the inversion of depth of 20cm and the measured soil moisture content is 0.6009. Through the analysis of accuracy and relative error, the inversion effect of 10cm ~ (20) cm is the most ideal. The average inversion accuracy of 10cm is 91.25 ~ (20) cm and the average of inversion accuracy is 88.53, which is 6.96% and 4.24% higher than the average value of 30cm inversion precision respectively. In the relative error analysis, the average relative error of 10 cm ~ (20) cm is lower than that of 30cm by 8.9% and 5.34 ~ (30) cm, respectively. The average relative error of 10 cm ~ (20) cm is 18.49. In conclusion, in drought monitoring, the effect of drought monitoring with depth less than 20cm is ideal when remote sensing and vertical drought index model are used to monitor drought.
【学位授予单位】：沈阳农业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：S423

【相似文献】