<![CDATA[The majority of farmers in Central Java, especially Sragen, grow rice. But many obstacles they face, especially the problem of diseases that attack rice plants. 5 types of rice diseases that are often found, namely leaf spot, blast, leaf blight, fronds, and tungro. Lack of understanding causes errors in handling, can cause harm. The purpose of this study is to improve accuracy in identifying diseases of rice plants affected by the disease. Identification of types of diseases in rice plants by combining the characteristics of shape and texture characteristics is important to improve the accuracy of the results. The method used influences the accuracy of each pattern in the image of rice plants. The data used are 70 training data and 30 test data with a size of 256 x 256 pixels. A combination of shape and texture characteristics is used to optimize accuracy. To get the characteristics of the segmentation process using Otsu and Morphology. The results are processed to obtain the shape characteristics using area and perimeter, and the characteristics of entropy, energy, homogeneity, correlation, and contrast texture, each of the 4 angles of GLCM, so that the features used are 22 features. This system uses a backpropagation method to classify types of diseases. The results obtained from 70 training data are 100% accuracy and 30 test data with an accuracy level of 93%]]>
Copyrights © 2020
