EKSTRAKSI KANDUNGAN AIR KANOPI DAUN TANAMAN PADI DENGAN DATA HYPERSPECTRAL

Agus Wibowo, Dian Ratnasari, Bangun Muljo Sukojo, Teguh Harianto, Yusuf S. Djajadihardja

Abstract


The spectral reflectance of leaves from several paddy stumps in irrigated rice area of Indramayu - West Java Province was measured with field spectrometer which has wavelength range 350-2500 nm and spectral resolution of 1 nm. The measurement was performed from a distance of 10 cm (FS10) and 50 cm (FS50) of rice leaf canopy surface. Measurement point is selected such that represent different stages of rice growth i.e. vegetative, reproductive and ripening. Meanwhile, a airborne survey with hyperpectral mapper (HYMAP) sensor is done over the same area. The HYMAP sensor has wavelength length from 400-2500nm, 128 channels, 10-20 nm spectral resolution, and 4.5 m spatial resolution. Clumps of rice that has been measured withdrawn and separated into the leaf, stem and panicle. Those sample are weighed as fresh weight and then dried in the laboratory until the temperature of 60 ° C. The canopy water content is the difference between fresh weight and dry weight. Linear regression modeling is performed to find the relationship between canopy water content (CWC) with water index (R900/R970) and the normalised difference water index (R860-R1240)/(R860-R1240). To obtain a good model of cross-validation process is carried out by separating 30% of data randomly for validation and the remaining 70% is used to build the regression model. Furthermore, the resulting model is validated with 30% of data, good model selected by the criteria of a maximum of R2 model and the RMSE of cross validation (RMSEVC) minimum. Selected regression model was applied to the data HYMAP to create distribution maps of canopy water content. Results show that both spectral indices shows a positive correlation and also sensitive to changes in water content that make the detection of canopy water content is possible. Selected model that can be used are (1) CWC = 13078.0 * WI - 13183.00 (R2 = 0:58, RMSECV = 499.13) and (2) CWC = 7157.4 * NDWI + 396.26 (R2 = 0:51, RMSECV = 640.84). Selected model is implemented with HYMAP data to produce a canopy water content distribution map. The result shows that both spectra indices has a negative value that represent a dry land. High spectral indices value also identified in a vegetative growth stage which is sparse rice canopy and mixed with water or soil background. Wider implementation should be done cautiously and need improvements to accommodate the level of the leaf canopy cover.

Keywords: Remote Sensing, Hyperspectral, Paddy/Rice, Canopy Water Content, HYMAP, Field Spectrometer.


Full Text:

PDF

Article Metrics

Abstract view : 194 times
PDF - 217 times

Refbacks

  • There are currently no refbacks.


Copyright (c) 2015 JURNAL ILMIAH GEOMATIKA

Geomatika Indexed by:

 

Copyright of Geomatika