Klasifikasi Berbagai Jenis Jamur Layak Konsumsi dengan Metode Backpropagation

Abstract

Mushrooms are a potential source of vegetable food around us. But there are difficulties in recognizing different types of mushrooms worthy of consumption because they have similarities when viewed visually. This study aims to determine the accuracy of the backpropagation method in classifying various types of mushrooms suitable for consumption automatically. The steps taken begin with data capture, initial processing, feature extraction and classification. The data used are images of six types of mushrooms worthy of consumption including button mushrooms, ears, straws, portabella, shitake, and ash oysters. Total data used 222 mushroom data, each type consists of 37 image data. The data sharing is 216 data as training data and testing data, and 6 data are used for single data test. The initial processing stages are grayscalling, adjustment, binaryization, noise reduction and resizing. Feature extraction is done to obtain color characteristics (RGB and HSI) and texture characteristics (texture statistics and gray level co-occurrence matrices). The test results obtained an accuracy of 97%.