IDENTIFIKASI CITRA KAYU JATI DAN KAYU AKASIA MENGGUNAKAN GLCM DAN ALGORITMA PERCEPTRON

MUHAMMAD MURAH PAMUJI, NIM. 12650033 (2017) IDENTIFIKASI CITRA KAYU JATI DAN KAYU AKASIA MENGGUNAKAN GLCM DAN ALGORITMA PERCEPTRON. Skripsi thesis, UIN SUNAN KALIJAGA YOGYAKARTA.

Preview	Text (IDENTIFIKASI CITRA KAYU JATI DAN KAYU AKASIA MENGGUNAKAN GLCM DAN ALGORITMA PERCEPTRON) 12650033_BAB-I_IV-atau-V_DAFTAR-PUSTAKA.pdf Download (3MB) | Preview
	Text (IDENTIFIKASI CITRA KAYU JATI DAN KAYU AKASIA MENGGUNAKAN GLCM DAN ALGORITMA PERCEPTRON) 12650033_BAB-II_sampai_SEBELUM-BAB-TERAKHIR.pdf Restricted to Registered users only Download (2MB)

Abstract

This essay discusses the identification of acacia and teak wood using feature extraction with Perceptron neural network method. Manual timber sorting is easily to be done by experts through the introduction of characteristic of each type of wood. Acacia and teak both are timberswhich have similar texture. Acacia wood has a sharper contrast than teak wood, this study will conduct testing using grayscale image. This study will use Gray Level Coocurence Matrix feature of grayscaled images. Features Gray Level Coocurence Matrixfeature which is used in this study are contrast, correlation, energy, and homogeneity. This study will use the values of these features for training and testing for Perceptron network. This study uses 35 data images for each type of woods that has been processed to be a plank. In the early stages, this research proceed with the image acquisition and pre-processing. Pre-processing performed is image resizing and grayscaling. The next step of this research is to perform feature extraction using Gray Level Coocurence Matrix on grayscaled images. This study uses combination of Gray Level Coocurence Matrix values. This study tested the combination of data matrix GLCM using artificial neural network perceptron. Based on test results, this research could yield high recognition accuracy of 95% on tests performed by using the correlation values.

Share this knowledge with your friends :

Actions (login required)

View Item