Tomato leaf disease detection using image processing technique

Abstract

Agriculture is the backbone of India. It contributes nearly 17% of the total GDP and it employs more than 60% population. Now a day’s plants are severely affected by a different type of diseases. The agriculture industry in all over the world are affected by severe economic losses due to these diseases in plants. In early days manual inspection was proposed to identify the disease, it is a difficult and time-consuming process and this can be overcome by automated methods in recent days. In this paper, a new method is proposed to identify disease in tomato plant. The proposed method consists of four stages, namely pre-processing, segmentation, feature extraction, and classification. In the pre-processing step, acquired images are resized and the noise was removed using the Weiner filter technique. Segmentation one of the important steps, for that modified K-Means image segmentation algorithm, was proposed. After the segmentation process, important features are extracted from the segmented image using Grey Level Co-occurrence Matrix (GLCM) feature extraction method. Finally, diseased leaves are classified using classification algorithms like Support Vector Machine (SVM) and Adaptive Neuro Fuzzy Inference System (ANFIS). The experiments are performed from the tomato leaf images in plant village datasets. The proposed methodology is tested for five types of diseases in tomato plants namely Bacterial Spot, Early Blight, Leaf Mold, Late Blight, Septoria Leaf Spot, and normal tomato leaf. Results proved that the classification accuracy was improved by using modified K-Means and ANFIS classifier when compared to the K-means segmentation with classification algorithms.