Identifikasi Daun Tanaman Jati Menggunakan Jaringan Saraf Tiruan Backpropagation dengan Ekstraksi Fitur Ciri Morfologi Daun

Abstract

Teak identification is very useful in life. Knowledge about kind of teak is very important for teak farmers to understand the characteristics, benefits, as well as the buying and selling price of teak. This research built an automatic teak identification system using backpropagation Neural Network with leaf morphology feature extraction. The five basic leaf characteristics that were used in this research are area, perimeter, diameter, length, and width of the leaf. The five basic characteristics were then derived into 12 derivative characteristics. The data were the images of teak leaves of Biotrop, Emas, Jobika, Muna, Prima, and Super. Each data consists of 20 images with the size of 1200 x 2300 pixel. In order to obtain the best accuracy, k-fold cross validation with k=5 was used. The result shows that the best accuracy was 84.17% with 17 hidden neuron of neural network