A sampling strategy based on NDVI prior knowledge for LAI ground measurements

Abstract

We propose a new sampling strategy based on Normalized Difference Vegetation Index(NDVI) prior knowledge for Leaf Area Index (LAI) ground measurements of non-homogeneous pixels. The method accuracy and stability have been analyzed in cases of different vegetation types and different pixel heterogeneity. The analysis results show that the proposed method is capable of properly dividing the non-homogeneous area into zones with different vegetation cover levels. It performed more accurate and robust than random sampling, systematic sampling and sampling based on classification in grassland and forest areas. The good performance indicates that this new sampling strategy for the LAI ground measurements may be used to remote sensing product validation for the heterogeneous pixels.