Klasifikasi Lahan Hijau dan Bangunan Menggunakan Data Citra Time Series EVI Landsat (Studi Area Tangerang, Bogor, dan Depok Tahun 2004)

Abstract

The increasing human population in the world has become drastically influence on urban land. Most of the urban area consisting of green land and buildings. Therefore, do a classification of green land area and the building uses image data Time Series of EVI Landsat in the study area of Tangerang, Bogor and Depok. The goal is to determine a pixel of the image is a green land or buildings; extracting and classifying characteristic of green land or buildings on the image; and perform comparisons process smoothing with non smoothing. In the classification of the study area of the city used a remote sensing system that produces a Landsat 7 satellite images to monitor the area of green land and buildings. In classifying the area of green land and buildings used a Time Series EVI Landsat Data. The method used is the Transformation Daubechies Wavelet one dimensional and Naive Bayes. Classification is done by evaluating the non-smoothing, smoothing level 1, level 2 smoothing, smoothing level 3, and smoothing level 4. The results of the classification by evaluation smoothing level 4 has the fewest number of errors and the highest percentage of correctness accuracy 90%. It can be concluded as a result of classification is best compared with the four other classification results.

Key words

classification, Time Series EVI, Transformasi Wavelet Daubechies, Naive Bayes, smoothing.