Leaf Area Index derivation from hyperspectral vegetation indicesand the red edge position

R. Darvishzadeh*, C. Atzberger, A. K. Skidmore, A. A. Abkar

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

112 Citations (Scopus)


The aim of this study was to compare the performance of various narrowband vegetation indices in estimating Leaf Area Index (LAI) of structurally different plant species having different soil backgrounds and leaf optical properties. The study uses a dataset collected during a controlled laboratory experiment. Leaf area indices were destructively acquired for four species with different leaf size and shape. Six widely used vegetation indices were investigated. Narrowband vegetation indices involved all possible two band combinations which were used for calculating RVI, NDVI, PVI, TSAVI and SAVI2. The red edge inflection point (REIP) was computed using three different techniques. Linear regression models as well as an exponential model were used to establish relationships. REIP determined using any of the three methods was generally not sensitive to variations in LAI (R2 < 0.1). However, LAI was estimated with reasonable accuracy from red/near-infrared based narrowband indices. We observed a significant relationship between LAI and SAVI2 (R2 = 0.77, RMSE = 0.59 (cross validated)). Our results confirmed that bands from the SWIR region contain relevant information for LAI estimation. The study verified that within the range of LAI studied (0.3 ≤ LAI ≤ 6.1), linear relationships exist between LAI and the selected narrowband indices.

Original languageEnglish
Pages (from-to)6199-6218
Number of pages20
JournalInternational Journal of Remote Sensing
Issue number23
Publication statusPublished - 2009
Externally publishedYes


Dive into the research topics of 'Leaf Area Index derivation from hyperspectral vegetation indicesand the red edge position'. Together they form a unique fingerprint.

Cite this