PREDIKSI PARAMETER-PARAMETER BIOFISIK TANAMAN PADI DARI DATA GROunDSPECTROmETER DAN HYPERSPECTRAL PESAWAT TERBANG DENGAN MENGGUNAKAN TEKNIK PARTIAL LEAST SquARE REGRESSIOn (PLSR) DAN nORmALIzED DIFFEREnCE SPECTRAL InDEx (NDSI)

Abstract

Paddy rice canopy hyperspectral was measured by using ground-based spectroradiometer and HyMap sensor onboard an airplane from 350 nm up to 2500 nm in the same time
that covered various growth stages. Coinciding with hyperspectral measurement, biophysical parameters such as leaf area index (LAI), SPAD value were measured on
ground during the airplane passed over area of interest (AOI). Rice yield was measured at the time of final harvesting by random selected yield (ubinan method) for each sampling area. In finding the best correlation among canopy hyperspectral reflectance with crop development, optimal individual waveband explored with involving all possible waveband combinations to obtain the best fitted two-pair waveband related to crop biophysical parameters. Normalized Difference Spectral Index (NDSI) was applied
from spectral transformations (obtained from optimal waveband selected by exploring all possible waveband) to improve sensitivity analysis on plant. Canopy hyperspectral
provided information about plant, soil and water background when plant canopy don’t completely cover soil surface yet. The present study was directed to examine spectral
indices and establish the relationships between biophysical parameters of rice by using partial least square regression (PLSR) technique.