CLUSTERING CITRA TANAH MENGGUNAKAN ALGORITMA FUZZY C-MEANS UNTUK MENILAI KESESUAIAN LAHAN PADA TANAMAN CENGKEH

DANANG AJI BIMANTORO, NIM. 13651060 (2017) CLUSTERING CITRA TANAH MENGGUNAKAN ALGORITMA FUZZY C-MEANS UNTUK MENILAI KESESUAIAN LAHAN PADA TANAMAN CENGKEH. Skripsi thesis, UIN SUNAN KALIJAGA YOGYAKARTA.

Preview	Text (CLUSTERING CITRA TANAH MENGGUNAKAN ALGORITMA FUZZY C-MEANS UNTUK MENILAI KESESUAIAN LAHAN PADA TANAMAN CENGKEH) 13651060_BAB-I_IV-atau-V_DAFTAR-PUSTAKA.pdf Download (4MB) | Preview
	Text (CLUSTERING CITRA TANAH MENGGUNAKAN ALGORITMA FUZZY C-MEANS UNTUK MENILAI KESESUAIAN LAHAN PADA TANAMAN CENGKEH) 13651060_BAB-II_sampai_SEBELUM-BAB-TERAKHIR.pdf Restricted to Registered users only Download (6MB)

Abstract

Clove is the one of commodities that have high economic value. This make many people want to plant cloves as much as possible, but there are many difficulties to plant a clove. This difficulties happens because when planting cloves there are several parameters that must be met, one of them is soil conditions. Suitable soil or not can be distinguished by color and texture, but many people do not know about it. Therefore, reasearchers will conduct research on soil image clustering based on color and texture features result of information gain feature selection using Fuzzy C-Means algorithm combined with genetic algorithm. The first step, the researches is performed image acquisition followed by a pre-processing. pre-processing results will be RGB color feature extraction and feature extraction texture of the first order order and continued to feature selection using Information gain to get the best features. Selected feature used to do clustering using Fuzzy C-Means algorithm combined with genetic algorithm. The next step is doing some testing followed by analysis phase to get the best result. Based on test result, it can be concluded that the Fuzzy C-Means algorithm combined with genetic algorithm provides good accuracy in image clustering process for suitable and unsuitable soil to the cloves plant based feature selection results using information gain. Best accuracy obtained from the use power of numbers 2 and selection of features with a threshold of 0.7 (mean red, mean green mean, variance, kurtosis) yields an accuracy of 88%.

Share this knowledge with your friends :

Actions (login required)

View Item