Grape Leaf Disease Diagnosis using Convolutional Neural Network and Support Vector Machines

The agricultural industry plays an important role in Thailand as it has an impact on the domestic economy and exportation. Well timed control and management of plant diseases are essential to help the quality control of products and reduce the damage of plant diseases or controlling appropriate chemicals amount possibly affects soil and water quality in the surrounding environment. This research presents a hybrid deep learning model for grape leaf disease diagnosis using a convolutional neural network (CNN) and support vector machine (SVM). This algorithm can be called Convolutional Support Vector Machine (cSVM). The deep learning algorithm is used in feature extraction part to increase the ability to extract distinct features using the CNN’s filters model, in which the user does not have to choose the parameters for testing by himself. The Grape leaf is tested for the diagnosis of 4 grape leaf cases including Healthy, Downey, Rust, and Scap. The performance of proposed algorithm showed the accuracy value equals 94.1 percent in average. Thus, the proposed algorithm can efficiently identify the infection types of the disease.