一种基于TVDI模型的煤田温度异常区面积提取方法研究

发布时间：2018-03-28 02:33

  本文选题：Landsat　切入点：温度植被干旱指数　出处：《干旱区地理》2017年03期

【摘要】：选用Landsat 8资料构建了地表温度(Ts)与归一化植被指数(NDVI)的Ts-NDVI特征空间,计算了温度植被干旱指数(TVDI)。利用MODIS温度产品数据和实地野外采样数据进行精度验证确定煤田火区TVDI阈值。通过对遥感影像的地表热异常信息进行定性与定量分析继而对煤田温度异常区边界信息进行挖掘。结果表明:(1)利用野外实测土壤相对含水量进行验证,反演值与实测值的相关系数R~2=0.66,表明干旱指数的反演精度较高,相关性较好。(2)TVDI模型对温度呈现出较高的敏感性,二维散点图集中在1∶1线上,对NDVI的敏感性较低,有利于识别温度异常区。(3)利用MVC最大合成法,建立TVDI-MVC作为精度验证数据,火区面积为5.03 km~2,其中TVDI-SC提取火区精度最大为98.50%,TVDI-SW_2提取火区精度最小为88.98%。可见煤田温度异常区范围较广,潜在的灾情恶化较严重。
[Abstract]:Using Landsat 8 data to build the land surface temperature (Ts) and normalized difference vegetation index (NDVI) of the Ts-NDVI feature space, the Temperature Vegetation Drought Index (TVDI) calculated by MODIS. The temperature of product data and field sampling data to verify the accuracy of determining coal fire TVDI threshold. The qualitative and quantitative analysis of the surface thermal anomaly of remote sensing information then the image boundary temperature anomaly information mining coal region. The results showed that: (1) relative water content are verified by the experimental field of soil, the correlation coefficient of R~2=0.66 inversion and the measured value, high inversion precision, the drought index is better4. (2) the TVDI model of temperature shows a higher sensitivity. Scatter plot is concentrated in 1 to 1 line, low sensitivity to NDVI, to identify the temperature anomaly area. (3) using the MVC synthesis method to establish TVDI-MVC as data validation, The fire area is 5.03 km~2, the maximum of 98.50% TVDI-SC extraction precision fire, fire precision TVDI-SW_2 extraction 88.98%. minimum visible coal temperature anomaly in a wide range of potential disaster deterioration is serious.

【作者单位】： 新疆大学资源与环境科学学院;新疆大学绿洲生态教育部重点实验室;新疆智慧城市与环境建模普通高校重点实验室;
【基金】：国家科技支撑计划项目资助(2014BAC15B01) 国家自然科学基金项目(41561089) 2016新疆研究生创新项目(XJGRI2016013)
【分类号】：TD752
，

本文编号：1674394