KLASIFIKASI PERTUMBUHAN TANAMAN PADI MENGGUNAKAN MODIS SINTETIK

Abstract

Paddy is one of the daily needs in Indonesia, which has many farmlands spread all around region. This condition cause a problem to get an information about paddy’s growth in farmland recently. The problem not only because the farmland spread all around, but also paddy has 3 phases which difficult to classify. The problem can be solve by using remote sensing, because it can give data about rice growth distribution from satellite such as MODIS. Satellite’s data can give a description and visualization of rice growth in some farming areas. In study, Fast Minimum Covariance Determinant (FMCD) and Feasible Solution Algorithm (FSA) were applied for rice growth stage classification using MODIS data. Both of them search an estimator to classify data with different way. FMCD search an estimator by comparing subset with another subset, while FSA search estimator by comparing each data from each subset. In this experiment had done twice. The first testing shows that both of the methods can classify about 50%. The second testing shows that FMCD can classify about 82% and FSA about 81%. The result obtained from MODIS’s testing data show that FMCD can give a better visualization of rice growth distribution than FSA