Development of Application for Pests and Diseases of Corn Android Based System

The research objective is to produce an Android-based expert system software that is capable of detecting pests and diseases in corn plants and is useful in providing information about symptoms and its control through image processing. This expert system program is processed through digital signal processing which consists of four (4) main parts, namely preprocessing, color feature extraction, texture feature extraction, and classification. The color feature extraction method used is The Color Moment as color feature extraction and GLCM (Gray-Level CooOccurrence Matrix) as a texture feature extraction. The classification method in this system uses K-Means clustering by dividing images into 4 clusters based on the color and texture of image objects. Training data using Multi SVM (Support Vector Machine) method. The result of this software program is named Corn Expert System (CES) which is installed on the desktop and the Android Cellphone (HP). This CES system application begins with taking pictures of corn leaves that are attacked by pests and diseases using Android phones by farmers in cornfields and sent to the desktop that is operated by the operator at the Agricultural Extension Office. Data from the desktop processing is sent back to the farmer via an android phone. The results of the detection of this CES program for pests, leaf scrapers and rust disease, leaf spot, leaf blight, and froth blight, have an accuracy level of up to 90%.