Classification of Tomato Plant Diseases Through Leaf Using Gray-Level Co-occurrence Matrix and Color Moment with Convolutional Neural Network Methods

Abstract

Tomato is a fruit vegetable source of vitamins and minerals; in addition to being consumed as fresh fruit, it can also be processed into food industry raw materials such as fruit juices and sauces. However, due to various causes such as diseases, pest attacks, and unstable weather conditions, there is a decrease in the quality and quantity of production. In order to contribute to maintaining the productivity of tomato plants, the use of technology can be an alternative to be applied to the cultivation of tomato plants. This study applies image processing techniques to detect the texture of affected leaf using gray-level co-occurrence matrix (GLCM) extraction and color moment using convolutional neural network (CNN) method. Among the diseases that often occur in tomato leaf are late blight, Septoria spot, bacterial spot, target spot, early blight, leaf curl, spider mites, two spotted spider mites, and leaf mold. In this study, a combination of GLCM-color moment and CNN method was chosen because of its reliability in identifying and classifying plant diseases compared to only using CNN. In this study, we used a combination of GLCM color moments and CNN methods.