Identification of chrysanthemum flower based on color and flower form using sobel edge and artificial neural network

Abstract

Chrysanthemum is a popular ornamental plant in Indonesia because it has a variety of shapes and colors that attract a lot of attention. In addition to the trimmer, this flower can be used as a mosquito repellent and also absorbs dirty air pollution. Chrysanthemum flowers can be distinguished by color and shape with two types of ways to grow Chrysanthemum flowers that grow in clusters are called spray and also flowers that grow with no cluster or a flower per stalk. In addition, Chrysanthemum flowers have various color variations including pink, purple, yellow, white, orange, and red. The purpose of this study is to identify the type of Chrysanthemum flowers using extraction of Red, Green, Blue (RGB) color features and characteristic extraction using a Sobel edge detection, for his identification method using Artificial Neural Networks. Chrysanthemum data are white puma, Vania, Iranian, purple aster, and pink standard used in this research consisted of 100 training data and 50 tests data. The test was done with four tests that is identification with form extraction (Sobel) yielded accuracy value equal to 59.52%, identification with color and shape extraction yield accuracy value equal to 65.36%, identification with color extraction on flower and base flower plate yield value accuracy of 69.44%, and identification with color extraction yields an accuracy value of 75.35%.