Turmeric (Curcuma domestica) is one of the plants of the Zingiberaceae family which is widely planted in the garden and in plants. One of the obstacles in the cultivation of turmeric plants is the existence of disease disorders, which among them are caused by pathogens. This research uses pathogenic microscopic images to be analyzed based on the texture and shape, then identified using Backpropagation Artificial Neural Networks. Texture feature extraction aims to get entropy, correlation, energy, contrast, and homogeneity values, while feature feature extraction to get area, perimeter, metric, eccentricity, and compactness values. The data used amounted to 126 images from 3 types of pathogens with each class consisting of 42 images. The types of pathogens are Colletotrichumcapsici, Curvularia sp., And Phakopsora. This data is then divided into 78 training data and 48 test data. Trials are done by changing the number of hidden layers found on the network. The results showed an accuracy rate of 94% with 2 hidden layers on the network
Volume 12 | 06-Special Issue
Pages: 95-102
DOI: 10.5373/JARDCS/V12SP6/SP20201011