基于Landsat数据的条件植被温度指数升尺度转换方法

Up-scaling Transformation Methods for Vegetation Temperature Condition Index Retrieved from Landsat Data

以陕西省关中平原为研究区域,应用点扩散函数、混合像素面积权重法和中值像素变异权重法将基于Landsat卫星遥感数据反演的分辨率为30 m的条件植被温度指数(VTCI)干旱监测结果上推至930 m的干旱监测结果,并与Aqua MODIS数据反演的分辨率为930 m的VTCI干旱监测结果进行对比分析,以期为两种空间尺度的干旱监测结果的综合应用提供技术支持。以MODIS数据反演的VTCI为参考,应用相关系数、均方根误差、半变异函数的估计值和图像纹理特征等对尺度上推的VTCI进行评价。结果表明,点扩散函数和混合像素面积权重法的尺度上推效果均较好,而中值像素变异权重法的尺度上推效果较差,说明点扩散函数和混合像素面积权重法均适用于研究区域VTCI干旱监测结果的尺度转换,且点扩散函数的数据处理过程更为简单。典型样点VTCI的尺度上推结果表明,空间异质性越小,尺度上推的结果越好。

Vegetation temperature condition index（ VTCI） drought monitoring results retrieved from Landsat remotely sensed data（ 30 m） in Guanzhong Plain,China were spatially transformed to a scale of the Aqua MODIS resolution（ 930 m） by using point spread function,mixed pixel area weighting method and median pixel variability weighting method. The transformed VTCIs were compared with the ones retrieved from Aqua MODIS data for agreement analysis of the two drought monitoring results. Taking MODIS VTCIs as the ‘real＇droughts,correlation coefficients and root mean square errors between the up-scaled Landsat VTCIs and MODIS VTCIs,and the texture and semi-variances of the two VTCIs were applied to select the best transformation method. The results showed that the transformed VTCIs from the point spread function and the mixed pixel area weighting method were better than those from the median pixel variability weighting method,which indicates that the point spread function and the mixed pixel area weighting method were both suitable for transforming the retrieved VTCI drought monitoring results from Landsat remotely sensed data,and the data processing procedure of the point spread function was relatively simple. The transformed VTCIs in the selected sampling sites covered by winter wheat showed that the smaller the spatial heterogeneity,the higher the transformed accuracy.