Identification of Rice Disease Types Based on Digital Images Leaves Using Algorithm Support Vector Machine (SVM)

Abstract

Diseases that commonly attack rice plant are: bacterial leaf blight, brown spot, and leaf smut. The disease can lead to a significant decrease in the quality and quantity of agricultural products so that it can cause economic loss. Farmers usually find out that the rice plant has been affected by the disease when it already shows severe symptoms and has spread to various parts of the plant so it is too late to control. Another thing that causes the failure of disease management in rice plant is the lack of knowledge of farmers about the symptoms of the disease which causes the provision of inappropriate treatment actions caused by errors in identifying the type of disease that appears with almost the same physical symptoms. Therefore, we need a solution in the form of a model that is able to identify the types of disease in rice plant so that farmers including the general public can detect disease that attack rice plant accurately and quickly. This study aims to develop a model for identification of rice plant disease based on plant leaf images. The research stages are: collecting rice leaf image data, sharing data for training and testing data using cross validation techniques, implementing the SVM algorithm and evaluating the model using a confusion matrix and calculating the accuracy, precision and recall of the SVM algorithm. The results of the evaluation of this identification model indicate that SVM can identify leaf disease of rice plant with an accuracy value of 0.90 or 90%, a precision of 0.902 or 90% and a recall of 0.900 or 90%.