Exploring the relationship between red edge parameters and crop variables for precision agriculture

This paper presents the study of the relationships between crop variables and the red edge parameters extracted using the inverted Gaussian model. Variability of the red edge parameters induced by the variations of leaf and canopy model parameters was analyzed using PROSPECT and SAILH simulated spectra. The position and shape of the red edge are influenced mostly by leaf area index(LAI) and chlorophyll content, confounded by the other model parameters. Red edge parameters were also extracted from CASI(Compact Airborne Spectrographic Imager) multitemporal hyperspectral data, and related with various crop variables. The study shows that red edge parameters are indicative of many crop properties, and the first derivative at the inflection position correlates well with green LAI, crop height and leaf water content. An empirical equation was built from the simulated spectra to predict LAI from the first derivative at the inflection position, and was applied to CASI hyperspectral data for green LAI retrieval. For all the samples including wheat, corn and soybean, comparison between the predicted and measured LAI resulted in a determination coefficient (R ²) of 0.86, and an RMSE of 0.61