THE CLASSIFICATION OF LAND COVER THROUGH SATELLITE IMAGE SPOT WITH GRAY LEVEL CO-OCCURRENCE MATRIX METHOD AND CLASSIFICATION OF K-NEAREST NEIGHBOR
Syaeful Hilman Supratman (1), Rita Magdalena (2), Sofia Saidah (3)

School of Electrical Engineering, Telkom University (1)(2)(3)
Jalan Telekomunikasi 01, Bandung 40257, Indonesia
syaeful.hilman[at]gmail.com (1), ritamagdalena[at]telkomuniversity.ac.id (2), sofiasaidahsfi[at]telkomuniversity.ac.id (3)

Abstract

The land is a crucial part that cannot be separated from human life and other living things. Nowadays, the function of many areas of land has been changing. This bears the intention of writers to discussed the problem which occurs in the classification of land cover to assist the government in deciding the use of some land areas in Bogor. This research used the Gray Level Co-Occurrence Matrix (GLCM) Method as a feature extraction method and K-Nearest Neighbor (k-NN) method for the process of classifying. The test result shows that this method can generate the best accuracy 85,8% when using 0 orientation parameter, quantization level = 8, the feature of statistical is contrast, correlation, energy, homogeneity, entropy, and when we used Euclidean type of k-NN with k = 3.

Keywords: Land coverage; GLCM; K-NN; euclidean

Topic: Computer and Communication Engineering

Link: https://ifory.id/abstract-plain/fX6W9JZpAKR3

Web Format | Corresponding Author (Syaeful Hilman Supratman)