Systematic analysis of the LUT-based inversion of PROSAIL using full range hyperspectral data for the retrieval of leaf area index in view of the future EnMAP mission

The upcoming satellite mission EnMAP offers the highly relevant opportunity of retrieving information on the seasonal development of vegetation parameters on a regional scale based on hyperspectral data. This study aims to investigate the potential of retrieving leaf area index (LAI) information from hyperspectral images. The widely used PROSAIL model is applied to generate look-up-table (LUT) libraries, by which the model is inverted to derive LAI information. Different techniques for the LUT based inversion are tested, such as several cost functions, type and amount of artificial noise, number of considered solutions and type of averaging method. The optimal inversion procedure (laplace, median, 4% inverse multiplicative noise, 350 averages) is identified by validating the results against corresponding in-situ measurements (N = 330) of LAI, leading to robust results (R² = 0.65, RMSE = 0.64).