植被覆盖对光学影像反演土壤含水量的影响分析

Analysis of the influence of vegetation cover on the inversion of soil water content from optical images

为了获取植被覆盖对光学影像反演土壤含水量的影响,选用归一化植被指数(NDVI)表示植被覆盖度水平,以垂直干旱指数(PDI)或修正垂直干旱指数(MPDI)线性回归土壤含水量的光学影像反演方法为例,基于2018—2020年哨兵2号卫星影像,利用河北省地面土壤含水量实测数据,对不同植被覆盖水平下土壤含水量反演精度进行评价,探讨植被覆盖对土壤含水量反演精度的影响。结果表明:(1)在NDVI≤0.13时,PDI的反演精度略优于MPDI反演精度,MPDI对PDI的修正作用不明显;(2)在0.13<NDVI≤0.30时,PDI仍能够达到较好的精度,但MPDI可以明显提高精度;(3)当0.30<NDVI≤0.59时,PDI反演精度较差,修正后MPDI能够显著提升精度;(4)在NDVI>0.59时,修正后的MPDI也无法得到满足要求的精度。可得出结论:当NDVI≤0.59时,光学影像可反演得到精度尚可的土壤含水量反演结果,通过一定的植被影响修正方法,可改善反演精度;当NDVI>0.59时,即使采用修正方法,基于光学影像难以得到满足要求的精度。

In order to obtain the influence of vegetation cover on the inversion of soil water content from optical data,NDVI was chosen to represent the vegetation cover,and a linear regression of vertical drought index(PDI)or modified vertical drought index(MPDI)was used as an example for the inversion method of soil water content from optical images,based on Sentinel 2 data from 2018-2020,using ground truth soil water content monitoring data in Hebei Province,and different soil water content inversion accuracy was evaluated at different NDVI levels,and the influence of NDVI on soil water content inversion accuracy was explored.The results show that(1)when NDVI≤0.13,the inversion accuracy of PDI is slightly smaller than that of MPDI,and the correction effect of MPDI on PDI is not obvious;(2)when 0.13<NDVI≤0.30,PDI can still achieve good accuracy,but MPDI can significantly improve the accuracy;(3)when 0.30<NDVI≤0.59,PDI inversion accuracy is poor,and the corrected MPDI can significantly improve the accuracy to meet the demand;(4)when NDVI>0.59,the corrected MPDI cannot get satisfactory accuracy either.